WO2019140703A1 - 一种用户画像的生成方法及装置 - Google Patents

一种用户画像的生成方法及装置 Download PDF

Info

Publication number
WO2019140703A1
WO2019140703A1 PCT/CN2018/073673 CN2018073673W WO2019140703A1 WO 2019140703 A1 WO2019140703 A1 WO 2019140703A1 CN 2018073673 W CN2018073673 W CN 2018073673W WO 2019140703 A1 WO2019140703 A1 WO 2019140703A1
Authority
WO
WIPO (PCT)
Prior art keywords
user
terminal
individual
group
image
Prior art date
Application number
PCT/CN2018/073673
Other languages
English (en)
French (fr)
Inventor
易晖
阙鑫地
张舒博
林于超
林嵩晧
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Priority to CN201880019023.3A priority Critical patent/CN110431585B/zh
Priority to PCT/CN2018/073673 priority patent/WO2019140703A1/zh
Publication of WO2019140703A1 publication Critical patent/WO2019140703A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising

Definitions

  • the embodiments of the present invention relate to the field of intelligent technologies, and in particular, to a method and an apparatus for generating a user image.
  • a terminal such as a mobile phone can abstract an actual user into a user portrait having one or more tags according to the user's usage behavior. For example, user A often uses a mobile phone to watch anime after 12 o'clock in the evening. Then, the mobile phone can use a label such as "late sleep” and "secondary" as the user image of user A. Subsequently, the mobile phone can provide customized services and functions for the user based on the user image of the user A, so as to improve the working efficiency of the mobile phone.
  • a user's complete user portrait usually contains multiple tags, some of which are individual tags generated directly based on the user's individual usage behavior on the mobile phone. For example, the above-mentioned "late sleep" tag is based on user A playing the mobile phone. Time generated.
  • some of the above tags are group tags that need to be obtained by big data calculation and data mining for different users' usage behaviors. For example, the server can determine that user A belongs to "90s" by comparing user images of multiple users. This group label.
  • the terminal needs to upload the user's behavior data to the server, but many behavior data related to the user's privacy cannot be uploaded to the server, resulting in a decrease in the accuracy of the user portrait generated by the server.
  • the terminal generates a user portrait for the user, since the terminal can only collect the behavior data of the single user using the terminal, the terminal cannot determine the group label to which the user belongs, and the accuracy of the generated user portrait is also caused. decline.
  • the embodiment of the present application provides a method and an apparatus for generating a user portrait, which can improve the accuracy of a user portrait generated by the terminal.
  • an embodiment of the present application provides a method for generating a user image, including: sending, by a terminal, at least one individual tag generated for a user to an image server, the individual tag reflecting a personal behavior characteristic of the user; At least one group label generated by the server for the user (the group label reflects behavior characteristics of the group to which the user belongs), the group label is generated by the image server based on the at least one individual label; thus, the terminal can use the group label to update A user's user portrait, thereby providing at least a portion of the updated user portrait to the first application.
  • the updated user portrait not only reflects the individual behavior characteristics of the user, but also reflects the group behavior characteristics of the user group, so the updated user image has higher integrity and accuracy, so that the first The application service provided by the application using the updated user portrait is more intelligent and accurate.
  • the method before the terminal sends the at least one individual tag generated by the user to the image server, the method further includes: collecting, by the terminal, behavior data generated when the user uses the terminal; the terminal generates the user data according to the behavior data. At least one individual tag, each individual tag including a type of the individual tag and a feature value of the individual tag.
  • the terminal sends the at least one individual tag generated by the user to the image server, including: the terminal sending the individual tag of the at least one individual tag whose sensitivity is less than the threshold to the image server, and the sensitivity is used by the terminal. Indicating the degree of correlation between the individual tag and the user's privacy, thereby reducing the risk of user privacy leakage while improving the accuracy of the user's portrait.
  • the terminal uses the group tag to update the user image of the user, including: the terminal adds the group tag to the user image of the user, and obtains an updated user image, and the updated user image
  • the group tag and at least one individual tag generated by the terminal are included.
  • the terminal uses the group label to update the user image of the user, including: the terminal updating the at least one individual label generated by the terminal according to the behavior data and the group label to obtain an updated user portrait, the update The updated user profile includes the updated individual tag.
  • the updated user image further includes the group tag.
  • the method further includes: the terminal receiving the first group label generated by the image server for the user and the first individual label Correlation degree, the first group label is one of the at least one group label, and the first individual label is one of the at least one individual label; the terminal corrects the feature value of the first individual label according to the degree of association, thereby further improving The accuracy of the user image generated by the subsequent terminal.
  • the terminal corrects the feature value of the first individual tag according to the degree of association, and specifically includes: the terminal, the sum of the feature value of the first individual tag and the correction value is used as the feature value corrected by the first individual tag, where
  • the correction value is a product of the degree of association and a preset correction factor, and the correction factor is used to reflect the degree of influence of the degree of association on the first individual label.
  • an embodiment of the present application provides a method for generating a user image, including: the image server acquiring an individual tag of at least one user; and the image server generating a group of the target user according to the individual tag of each of the at least one user a tag, the group tag reflects a behavior characteristic of the group to which the target user belongs, the target user is one of the at least one user; the portrait server sends the group tag of the target user to the terminal.
  • the image server generates a group tag of the target user according to the individual tags of each of the at least one user, including: the image server according to the individual tags of each of the at least one user, At least one user is divided into at least one group; the portrait server uses the label of the group to which the target user belongs as the group label of the target user.
  • the image server divides the at least one user into at least one group according to the individual tags of each of the at least one user, including: the image server is based on the individual of each of the at least one user
  • the tag divides the at least one user into at least one group by one or more of clustering, feature combination post-clustering, and feature-converted clustering.
  • the method further includes: the image server determining the degree of association between the group tags of the target users and each individual tag of the target user; The server sends the association degree to the terminal.
  • an embodiment of the present application provides a terminal, including: a portrait management module, and a data collection module, a portrait calculation module, a portrait optimization module, and an image query module connected to the portrait management module, wherein the portrait a management module, configured to: send at least one individual tag generated for the user to the portrait server, the individual tag reflects a personal behavior characteristic of the user; and receive at least one group tag generated by the image server for the user, the group tag is an image The server is generated based on the at least one individual tag, the group tag reflects a behavior characteristic of the group to which the user belongs; the portrait optimization module is configured to: update the user image of the user by using the group tag; the image query module is configured to: : providing at least a portion of the updated user portrait to the first application.
  • the data collection module is configured to: collect behavior data generated when the user uses the terminal; the image calculation module is configured to: generate at least one individual label for the user according to the behavior data, each The individual tag includes the type of the individual tag and the feature value of the individual tag.
  • the portrait management module is configured to: send the individual label of the at least one individual label whose sensitivity is less than a threshold to the image server, where the sensitivity is used to indicate the individual label and the user's privacy. The degree of correlation between them.
  • the image optimization module is specifically configured to: add the group label to the user image of the user, to obtain an updated user portrait, and the updated user image includes the group label and At least one individual tag generated by the terminal.
  • the image optimization module is configured to: update the at least one individual tag generated by the terminal according to the behavior data and the group tag to obtain an updated user image, where the updated user image includes The updated individual label.
  • the updated user image also includes the group tag.
  • the image management module is further configured to: receive a degree of association between the first group tag generated by the image server for the user and the first individual tag, where the first group tag is the at least one One of the group tags, the first individual tag is one of the at least one individual tags; the portrait optimization module is further configured to: correct the feature value of the first individual tag according to the degree of association.
  • the image optimization module is specifically configured to: use the sum of the feature value of the first individual tag and the correction value as the feature value after the first individual tag is corrected, and the correction value is the correlation.
  • the product of the degree and the preset correction factor which is used to reflect the degree of influence of the degree of association on the first individual label.
  • an embodiment of the present application provides a server, including a portrait management module, and a portrait calculation module connected to the portrait management module, wherein the portrait management module is configured to: acquire an individual label of at least one user;
  • the image calculation module is configured to: generate, according to an individual label of each user of the at least one user, a group label of the target user, where the group label reflects behavior characteristics of the group to which the target user belongs, and the target user is the at least one user
  • the image management module is further configured to: send the group label of the target user to the terminal.
  • the image calculation module is configured to: divide the at least one user into at least one group according to an individual label of each user of the at least one user; and label the group to which the target user belongs As the group tag for the target user.
  • the image calculation module is specifically configured to: perform one of clustering, feature combination post-clustering, and feature-switching clustering based on individual tags of each user of the at least one user Or dividing the at least one user into at least one group in multiple ways.
  • the image calculation module is further configured to: determine a degree of association between the group label of the target user and each individual label of the target user; the image management module is further configured to: The degree of association is sent to the terminal.
  • an embodiment of the present application provides a terminal, including: a processor, a memory, a bus, and a communication interface; the memory is configured to store a computer execution instruction, and the processor is connected to the memory through the bus, when the terminal is running The processor executes the computer-executed instructions stored in the memory to cause the terminal to execute the method of generating any of the user portraits described above.
  • an embodiment of the present application provides an image server, including: a processor, a memory, a bus, and a communication interface; the memory is configured to store a computer execution instruction, and the processor is connected to the memory through the bus, and the image server In operation, the processor executes the computer execution instructions stored in the memory to cause the image server to execute the method of generating any of the user images described above.
  • the embodiment of the present application provides a computer readable storage medium, where the computer readable storage medium stores an instruction, when the instruction is run on any of the foregoing terminals, causing the terminal to execute any one of the user images The method of generation.
  • an embodiment of the present application provides a computer readable storage medium, where the computer readable storage medium stores an instruction, when the instruction is run on any one of the image servers, causing the image server to execute any of the above The method of generating user images.
  • the embodiment of the present application provides a computer program product including instructions, when the terminal runs on any of the above terminals, causing the terminal to execute the method for generating the user portrait.
  • the embodiment of the present application provides a computer program product including instructions, when the image server is run on any of the image servers, to cause the image server to execute the method for generating the user image.
  • the names of the components in the terminal or the image server are not limited to the device itself, and in actual implementation, the components may appear under other names. As long as the functions of the various components are similar to the embodiments of the present application, they are within the scope of the claims and their equivalents.
  • FIG. 1 is a schematic structural diagram 1 of a terminal according to an embodiment of the present disclosure
  • FIG. 2 is a schematic structural diagram of a user portrait platform provided by an embodiment of the present application.
  • FIG. 3 is a schematic structural diagram 1 of a user portrait module according to an embodiment of the present application.
  • FIG. 5 is a schematic structural diagram 2 of a user portrait module according to an embodiment of the present disclosure.
  • FIG. 6 is a schematic diagram of a user label according to an embodiment of the present application.
  • FIG. 7 is a schematic structural diagram 3 of a user portrait module according to an embodiment of the present disclosure.
  • FIG. 8 is a schematic structural diagram of an image server according to an embodiment of the present application.
  • 9A is a schematic diagram 1 of a principle for generating a group label according to an embodiment of the present application.
  • 9B is a schematic diagram 2 of a principle for generating a group label according to an embodiment of the present application.
  • FIG. 10 is a schematic diagram 3 of a principle for generating a group label according to an embodiment of the present application.
  • FIG. 11 is a schematic flowchart diagram of a method for generating a user image according to an embodiment of the present application.
  • FIG. 12 is a schematic diagram of a method for generating a user portrait according to an embodiment of the present application.
  • FIG. 13 is a schematic structural diagram 2 of a terminal according to an embodiment of the present disclosure.
  • FIG. 14 is a schematic structural diagram of an image server according to an embodiment of the present application.
  • some intelligent reminders or services can be performed on the terminal based on the user's historical behavior habits or based on some rules or models, so that the user can more conveniently use the terminal, making the user feel that the terminal is more and more intelligent. Chemical.
  • the terminal can implement various intelligent services by itself or by combining with the cloud.
  • the terminal may include a rule platform, an algorithm platform, and a user portrait module.
  • the terminal can implement various intelligent services through one or more of the three platforms and other resources, for example: 1. service recommendation service; 2. reminding service; 3. notification filtering service.
  • the terminal includes a recommendation service framework for implementing the service recommendation service, and the recommendation service framework may at least include an algorithm platform, a rule platform, and a user portrait module.
  • the rule platform can match the service that the user of the terminal wishes to use in the current scenario according to the rule.
  • the above algorithm platform can predict the service that the user of the terminal wishes to use in the current scenario according to the model.
  • the recommendation service framework may place the service predicted by the rule platform or the algorithm platform in a display interface of the recommended application, so that the user can conveniently enter the interface corresponding to the service through the display interface of the recommended application.
  • the above rules can be sent to the terminal by the server (that is, the cloud).
  • the rule can be obtained by big data statistics or by empirical data.
  • the above model can be obtained by training user history data and user feature data through the algorithm platform to obtain a model. And the model can be updated based on new user data and feature data.
  • the user history data may be behavior data of the terminal used by the user for a period of time.
  • the user profile data may include a user profile or other type of feature data, which may be, for example, behavior data of the current user.
  • the user portrait can be obtained through the user portrait module in the terminal.
  • the terminal includes a recommendation framework for implementing the reminder service.
  • the recommendation framework can include at least a rules platform, a graphical user interface, and a user portrait module.
  • the above rule platform can listen to various events.
  • the application in the terminal can register various rules to the rule platform; then the rule platform listens to various events in the terminal according to the registered rules; matches the monitored event with the rule, and listens to the event and some
  • the reminder corresponding to the rule is triggered, that is, a highlight event is recommended to the user.
  • the reminder is ultimately displayed by the graphical user interface or by the application of the registration rule.
  • the condition of some rules may be a limitation on the user's portrait.
  • the rule platform may request the current user portrait from the user portrait module to determine whether the current user portrait matches the conditions in the rule.
  • the terminal includes a notification filtering framework for implementing the notification filtering service.
  • the notification filtering framework may include at least a rule platform, an algorithm platform, and a user portrait module.
  • the type of the notification may be determined by the rule platform, and the type of the notification may be determined by the algorithm platform. Then, based on the type of the notification and the preference of the user, it is determined whether the notification is a notification of interest to the user, and a different manner of reminder display is performed for the notification that the user is interested in and the notification that the user is not interested.
  • the user's preferences may include the user's portrait, as well as the user's historical processing behavior for certain types of notifications. Among them, the user portrait is provided by the user portrait module.
  • the terminal may include a rule platform that provides the capabilities required for each framework to the above three frameworks.
  • the terminal may also include a plurality of rule platforms, each of which provides capabilities to the above three frameworks.
  • the terminal may include an algorithm platform that provides the required capabilities of each framework to the recommended service framework and the notification filtering framework; or the terminal may also include two algorithm platforms to provide capabilities to the two frameworks respectively.
  • the terminal can include a user portrait module that provides the capabilities required for each framework to the three frameworks described above.
  • the terminal may also include a plurality of user portrait modules that provide capabilities to each of the frames.
  • the user portrait module provided by the embodiment of the present invention may be included in the terminal.
  • the terminal can be, for example, a mobile phone, a tablet personal computer, a laptop computer, a digital camera, a personal digital assistant (PDA), a navigation device, and a mobile internet device. , MID) or wearable device, etc.
  • FIG. 1 is a block diagram showing a partial structure of a terminal according to an embodiment of the present invention.
  • the terminal is described by taking the mobile phone 100 as an example.
  • the mobile phone 100 includes: a radio frequency (RF) circuit 110 , a power source 120 , a processor 130 , a memory 140 , an input unit 150 , a display unit 160 , a sensor 170 , and audio Circuit 180, and components such as wireless-fidelity (Wi-Fi) module 190.
  • RF radio frequency
  • the structure of the handset shown in FIG. 1 does not constitute a limitation to the handset, and may include more or less components than those illustrated, or some components may be combined, or different components may be arranged.
  • the components of the mobile phone 100 will be specifically described below with reference to FIG. 1 :
  • the RF circuit 110 can be used to send and receive information or to receive and transmit signals during a call.
  • the RF circuit 110 may send downlink data received from the base station to the processor 130 for processing, and send the uplink data to the base station.
  • RF circuits include, but are not limited to, RF chips, antennas, at least one amplifier, transceiver, coupler, low noise amplifier (LNA), duplexer, RF switch, and the like.
  • RF circuitry 110 can also communicate wirelessly with networks and other devices.
  • the wireless communication may use any communication standard or protocol, including but not limited to global system of mobile communication (GSM), general packet radio service (GPRS), code division multiple access (code) Division multiple access (CDMA), wideband code division multiple access (WCDMA), long term evolution (LTE), e-mail, short messaging service (SMS), and the like.
  • GSM global system of mobile communication
  • GPRS general packet radio service
  • code division multiple access code division multiple access
  • WCDMA wideband code division multiple access
  • LTE long term evolution
  • SMS short messaging service
  • the memory 140 can be used to store software programs and modules, and the processor 130 executes various functional applications and data processing of the mobile phone 100 by running software programs and modules stored in the memory 140.
  • the memory 140 may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application required for at least one function (such as a sound playing function, an image playing function, etc.), and the like; the storage data area may be stored according to The data created by the use of the mobile phone 100 (such as audio data, phone book, etc.) and the like.
  • the memory 140 may include a high speed random access memory, and may also include a nonvolatile memory such as at least one magnetic disk storage device, flash memory device, or other volatile solid state storage device.
  • the memory 140 can also store a knowledge base, a tag library, and an algorithm library.
  • the input unit 150 can be configured to receive input numeric or character information and to generate key signal inputs related to user settings and function control of the handset 100.
  • the input unit 150 may include a touch panel 151 and other input devices 152.
  • the touch panel 151 also referred to as a touch screen, can collect touch operations on or near the user (such as the user using a finger, a stylus, or the like on the touch panel 151 or near the touch panel 151. Operation), and drive the corresponding connecting device according to a preset program.
  • the touch panel 151 may include two parts: a touch detection device and a touch controller.
  • the touch detection device detects the touch orientation of the user, and detects a signal brought by the touch operation, and transmits the signal to the touch controller; the touch controller receives the touch information from the touch detection device, converts the touch information into contact coordinates, and sends the touch information.
  • the processor 130 is provided and can receive commands from the processor 130 and execute them.
  • the touch panel 151 can be implemented in various types such as resistive, capacitive, infrared, and surface acoustic waves.
  • the input unit 150 may also include other input devices 152.
  • other input devices 152 may include, but are not limited to, one or more of a physical keyboard, function keys (such as volume control buttons, switch buttons, etc.), trackballs, mice, joysticks, and the like.
  • the display unit 160 can be used to display information input by the user or information provided to the user and various menus of the mobile phone 100.
  • the display unit 160 may include a display panel 161.
  • the display panel 161 may be configured in the form of a liquid crystal display (LCD), an organic light-emitting diode (OLED), or the like.
  • the touch panel 151 can cover the display panel 161. When the touch panel 151 detects a touch operation on or near the touch panel 151, the touch panel 151 transmits to the processor 130 to determine the type of the touch event, and then the processor 130 according to the touch event. The type provides a corresponding visual output on display panel 161.
  • the touch panel 151 and the display panel 161 are two independent components to implement the input and input functions of the mobile phone 100 in FIG. 1, in some embodiments, the touch panel 151 may be integrated with the display panel 161. The input and output functions of the mobile phone 100 are implemented.
  • the handset 100 can also include at least one type of sensor 170, such as a light sensor, motion sensor, and other sensors.
  • the light sensor may include an ambient light sensor and a proximity sensor, wherein the ambient light sensor may adjust the brightness of the display panel 161 according to the brightness of the ambient light, and the proximity sensor may close the display panel 161 when the mobile phone 100 moves to the ear. / or backlight.
  • the accelerometer sensor can detect the magnitude of acceleration in all directions (usually three axes). When it is stationary, it can detect the magnitude and direction of gravity. It can be used to identify the gesture of the mobile phone (such as horizontal and vertical screen switching, related Game, magnetometer attitude calibration), vibration recognition related functions (such as pedometer, tapping).
  • the mobile phone 100 can also be configured with other sensors such as a gyroscope, a barometer, a hygrometer, a thermometer, an infrared sensor, and the like, and will not be described herein.
  • the audio circuit 180, the speaker 181, and the microphone 182 can provide an audio interface between the user and the handset 100.
  • the audio circuit 180 can transmit the converted electrical data of the received audio data to the speaker 181 for conversion to the sound signal output by the speaker 181; on the other hand, the microphone 182 converts the collected sound signal into an electrical signal by the audio circuit 180. After receiving, it is converted into audio data, and then the audio data is output to the RF circuit 110 for transmission to, for example, another mobile phone, or the audio data is output to the memory 140 for further processing.
  • Wi-Fi is a short-range wireless transmission technology.
  • the mobile phone 100 can help users to send and receive emails, browse web pages, and access streaming media through the Wi-Fi module 190, which provides users with wireless broadband Internet access.
  • FIG. 1 shows the Wi-Fi module 190, it can be understood that it does not belong to the essential configuration of the mobile phone 100, and may be omitted as needed within the scope of not changing the essence of the invention.
  • the processor 130 is the control center of the handset 100, which connects various portions of the entire handset using various interfaces and lines, by running or executing software programs and/or modules stored in the memory 140, and recalling data stored in the memory 140, The various functions and processing data of the mobile phone 100 are executed, thereby realizing various services based on the mobile phone.
  • the processor 130 may include one or more processing units; preferably, the processor 130 may integrate an application processor and a modem processor, where the application processor mainly processes an operating system, a user interface, an application, and the like.
  • the modem processor primarily handles wireless communications. It can be understood that the above modem processor may not be integrated into the processor 130.
  • the processor 130 may execute program instructions stored in the memory 140 to implement the method shown in the following embodiments.
  • the mobile phone 100 also includes a power source 120 (such as a battery) that supplies power to various components.
  • a power source 120 such as a battery
  • the power source can be logically coupled to the processor 130 through a power management system to manage functions such as charging, discharging, and power consumption through the power management system.
  • the mobile phone 100 may further include a camera, a Bluetooth module, and the like, which are not described herein.
  • the terminal provided by the embodiment of the present invention includes a user portrait module, and the user portrait module can abstract a user's information by collecting and analyzing various behavior data of the user who uses the terminal. According to the request of the application, the user portrait module can predict the current possible behavior or preference of the user through the abstracted information, and return the predicted result to the application, that is, return the user profile to the application.
  • the user portrait usually includes one or more user tags for reflecting user characteristics.
  • One user tag can be divided into two parts, one is a user tag type, and the other part is a feature value of the user tag.
  • user portrait of user A includes four user tags, wherein the type of user tag 1 is "gender", and the feature value of the user tag 1 is female, that is, user A is described.
  • the gender of the user is 2; the type of the user tag 2 is "home address", the feature value of the user tag 2 is Beijing, that is, the user A lives in Beijing; the type of the user tag 3 is "day and night”, and the user tag 3
  • the eigenvalue is "85 points" (with a perfect score of 100 points), which means that user A has a higher probability of generating day and night behavior.
  • user B also has a "day and night” type of user tag, if it is scored as "60 points”, it means that the probability that user B generates day and night behavior is less than the probability that user A generates day and night behavior.
  • the user tags in the user portrait can be divided into two types: individual tags and group tags.
  • the individual tag refers to a user feature that can be directly abstracted based on behavior data when the user uses the terminal, and the individual tag generally reflects the user's personal behavior feature. For example, the user often uses the mobile phone after 12 o'clock in the evening, and the mobile phone can determine the individual tag of the type of "day and night" and the characteristic value of the individual tag (ie, the scoring situation) according to the user's usage habits.
  • the group label refers to a label generated by the user according to the characteristics of the group to which the user belongs, after analyzing the behavior data of the plurality of users. For example, according to the behavior data of the user A, the user B, and the user C, it can be determined that the user A, the user B, and the user C all belong to a group having the characteristics of “day and night”, and the group label of the group includes “workaholics”, then, "Workaholic" is a group tag for User A, User B, and User C.
  • FIG. 2 is a schematic structural diagram of a user portrait platform according to an embodiment of the present invention.
  • the user portrait platform includes at least one terminal 10 and a portrait server 30, wherein the terminal 10 includes a user portrait module 20.
  • the user portrait module 20 described above can provide a user portrait for a variety of applications in the terminal 10.
  • the application can be a system level application or a general level application.
  • System-level applications generally refer to: The application has system-level permissions to access a variety of system resources.
  • a general-level application generally refers to the fact that the application has normal permissions, may not be able to obtain certain system resources, or requires user authorization to obtain some system resources.
  • the system level application can be an application pre-installed in the terminal 10.
  • the common level application can be an application pre-installed in the terminal 10, or an application installed by a subsequent user.
  • the user portrait module 20 can provide a user portrait to a system-level application such as a service recommendation application, a reminder application, and a notification filtering application, respectively.
  • the service recommendation application, the reminder application, and the notification filtering application are respectively used to implement the service recommendation service, the reminder service, and the notification filtering service in the foregoing embodiments.
  • the user portrait module 20 can also provide user portraits for video applications, news applications, or other applications.
  • the user portrait module 20 can also communicate with the image server 30 on the cloud side (ie, the network side).
  • the user portrait module 20 may send an individual tag of the generated user portrait that does not involve user privacy to the portrait server 30, thereby reducing the risk of user privacy leakage.
  • the portrait server 30 may combine the individual labels of other one or more users sent by other terminals 10, and determine the user by feature clustering, combination, or feature conversion. The group to which A belongs, thereby generating the group label of user A.
  • the portrait server 30 may send the group tag generated for the user A to the terminal 10, and the user portrait module 20 combines the individual tag of the user A with the group tag of the user A to generate a user image with higher integrity and accuracy for the user A. Thereby, the accuracy of the user portrait used by the terminal 10 is improved.
  • FIG. 3 is a schematic structural diagram of a user portrait module in the terminal 10 according to an embodiment of the present invention.
  • the user portrait module 20 may include a first portrait management module 201, a data collection module 202, a first portrait calculation module 203, a portrait optimization module 204, and an image query module 205.
  • the data collection module 202 provides the user portrait module 20 with collection capability support for the underlying metadata.
  • the data collection module 202 can collect behavior data generated when the user uses the terminal 10, and store and read and write the collected behavior data.
  • FIG. 4 is a schematic diagram of behavior data provided by an embodiment of the present invention.
  • the behavior data collected by the data collection module 202 may specifically include application level data 401, system level data 402, and sensor level data 403.
  • the application level data 401 may include data collected by the application layer at the runtime to reflect user behavior characteristics, such as an application name, an application usage time, a usage duration, and the like.
  • the data collection module 202 can also collect the video name being played, the video stop time, the number of video play sets, the total number of video sets, etc.; when the running application is a music application The data collection module 202 can also collect the name of the music being played, the type of music, the duration of the playing, the playing frequency, and the like; when the running application is a gourmet application, the data collecting module 202 can also collect the current store name, the type of the food, Store address, etc.
  • the data collection module 202 may also use the image text sensing technology to collect data according to specific situations, for example, identifying the text content in the image through optical character recognition (OCR) technology to obtain Text information in the picture.
  • OCR optical character recognition
  • System level data 402 may include data collected at runtime that various services provided in the framework may reflect user behavior characteristics.
  • the data collection module 202 can listen to a broadcast message from an operating system or an application, and obtain information such as a Bluetooth switch state, a SIM card state, an application running state, an automatic rotation switch state, a hotspot switch state, and the like through a monitoring service; for example, data collection.
  • the module 202 can obtain real-time scene information of the system by calling a specific interface, such as a contact provider API provided by the Android system, a content provider API, a calendar provider API, and the like. For example, audio, video, pictures, contacts, schedules, time, date, battery, network status, headset status, and more.
  • Sensor level data 403 may include data collected by devices such as sensors for reflecting user behavior characteristics.
  • data generated by sensors such as distance sensors, acceleration sensors, air pressure sensors, gravity sensors, or gyroscopes can be used to identify the user's behavioral states: onboard, cycling, walking, running, stationary, and others.
  • the collection period of the data collection module 202 can be set to an acquisition period with a short duration.
  • the collection period can be any value that does not exceed 24 hours.
  • the data collection module 202 can collect the GPS data of the terminal 10 every 5 minutes, and collect the number of images stored in the library in the terminal 10 every 24 hours. In this way, the terminal 10 only needs to maintain the behavior data of the user collected in the last 24 hours, and avoid occupying too many computing resources and storage resources of the terminal 10.
  • the data collection module 202 can collect the application level data 401, the system level data 402, and the sensor level data 403 by means of system monitoring, reading a specific data interface, invoking a system service, and collecting a collection.
  • the first image calculation module 203 may include a generation algorithm or model of a series of individual tags, and the first image calculation module 203 is configured to receive behavior data of the user collected by the data collection module 202 within a certain time, and according to the above algorithm or model. Determine the individual label of the user.
  • the first image management module 201 may send the behavior data collected by the data collection module 202 in the last 24 hours to the first image calculation module 203, and the first image calculation module 203 follows the above algorithm. Or the model determines a plurality of individual tags that reflect the characteristics of the user's behavior through statistical analysis, machine learning, and the like.
  • the first image calculation module 203 may also desensitize the individual tags related to user privacy, thereby reducing the sensitivity of the individual tags.
  • the user's individual tags include, but are not limited to, the following six types of tags: basic attributes, social attributes, behavioral habits, hobbies, psychological attributes, and mobile phone usage preferences.
  • the above basic attributes include but are not limited to: personal information and physiological characteristics.
  • the personal information includes, but is not limited to, name, age, document type, education, constellation, belief, marital status, and email address.
  • the residence of the house may include: renting a house, owning a house, and repaying the loan.
  • the mobile phone can include: a brand and a price.
  • the mobile operator may include: brand, network, traffic characteristics, and mobile number.
  • the brands may include: Mobile, China Unicom, telecommunications, and others.
  • the network may include: none, 2G, 3G, and 4G.
  • the flow characteristics may include: high, medium, and low.
  • the above behaviors include but are not limited to: geographical location, lifestyle, transportation, residential hotel type, economic/financial characteristics, dining habits, shopping characteristics and payment.
  • the living habits may include: work schedule, home time, work time, computer online time, and grocery shopping time.
  • the shopping characteristics may include: a shopping item category and a shopping method.
  • the payment situation may include: payment time, payment location, payment method, single payment amount, and total payment amount.
  • the above hobbies include but are not limited to: reading preferences, news preferences, video preferences, music preferences, sports preferences, and travel preferences.
  • the reading preferences may include: reading frequency, reading time period, total reading time, and reading classification.
  • the above psychological attributes include, but are not limited to, lifestyle, personality, and values.
  • the above mobile phone usage preferences include, but are not limited to, application preferences, notification alerts, in-app operations, user preferences, system applications, and common settings.
  • the first image management module 201 can combine the dynamic scenes of the current user, for example, current time, current position (latitude and longitude), motion state, weather, location. (POI), mobile phone status, and switch status, etc., to obtain a perception of the current real-time scene, for example, the perceived result is on the way to work, travel, and so on.
  • the terminal can predict the subsequent behavior of the user on the terminal, thereby providing an intelligent customized personalized service, for example, automatically displaying the home route and the road condition for the user during the off-hours of the user. .
  • the various individual labels described above are merely examples.
  • the specific individual label in the maintenance in the first image calculation module 203 may be expanded according to the requirements of the service, and a new type of label may be added, or a more detailed classification may be performed on the existing label.
  • the individual tags generated by an image calculation module 203 for the user may reflect the personalized features of the user.
  • the individual label image may be cached in a database (for example, SQLite) of the terminal of the terminal 10 for a certain period of time (for example, 7 days).
  • the label that does not involve user privacy in the individual tag may be transmitted to the portrait server 30 by the first portrait management module 201.
  • the terminal 10 may encrypt the above-mentioned individual tags by using a preset encryption algorithm, for example, an Advanced Encryption Standard (AES), and store the encrypted individual tags in the SQLite to improve the individual tags in the terminal 10.
  • AES Advanced Encryption Standard
  • the first image management module 201 is coupled to the data collection module 202, the first image calculation module 203, the image optimization module 204, and the image query module 205.
  • the first portrait management module 201 is a control center for providing a user portrait service in the terminal 10, and can be used for providing various management functions and running scripts of the user portrait service, for example, starting a service for establishing a user portrait, from the data collection module 202.
  • Obtaining behavior data of the user instructing the first portrait calculation module 203 to calculate the individual label of the user, instructing the portrait optimization module 204 to generate a complete user portrait including the individual label and the entire label of the user, and instructing the portrait query module 205 to identify the user identity
  • the terminal 10 may transmit the generated individual tag to the portrait server 30 based on the post/get request method in the hypertext transfer protocol over secure socket layer (HTTPS) protocol.
  • HTTPS secure socket layer
  • the personal privacy of the user is not revealed in the individual tags sent by the terminal 10 to the portrait server 30, and the subsequent image server 30 can determine the group tag to which the user belongs according to the received individual tags, thereby obtaining a complete and accurate user portrait.
  • the first image management module 201 can input the individual tags generated by the first image calculation module 203 and the group tags of the users sent by the image server 30 to the image optimization module 204.
  • the portrait optimization module 204 can use the group label as the new behavior data to generate a complete user image in combination with the original collected behavior data, and the user behavior model and the user belonging group are comprehensively considered when generating the user image.
  • the group behavior feature so the user image obtained by the portrait optimization module 204 includes both the individual label of the user and the group label of the user, so that the integrity and accuracy of the user portrait can be improved.
  • the portrait server 30 may further calculate the degree of association between the user's group tag and the individual tag.
  • the user's individual tags are "online shopping" and "game”
  • the image server 30 generates a user-generated group tag as "home”.
  • the image server 30 can further calculate the "home” group tag and the "online shopping” respectively.
  • the subsequent portrait optimization module 204 can also correct the feature values of the two individual tags "online shopping" and "game” according to the above-mentioned degree of association, thereby improving the accuracy of the finally generated user portrait.
  • the image query module 205 is configured to respond to a request for querying a user image by any application in the application layer.
  • the portrait query module 205 can provide a Provider interface of the Android unified standard, and the application can request the first portrait management module 201 to provide a user portrait to the Provider interface by calling the Provider interface.
  • the image query module 205 provides a user portrait to the application
  • the user identity requesting the user image can be authenticated by means of digital signature or the like to reduce the risk of user privacy leakage.
  • FIG. 8 is a schematic structural diagram of an image server according to an embodiment of the present invention.
  • the portrait server 30 may include a second portrait management module 301 and a second portrait calculation module 302.
  • Second portrait management module 301 Second portrait management module 301
  • the second portrait management module 301 is a control center that provides a user portrait service in the image server 30, and the second portrait management module 301 is connected to the second portrait calculation module 302.
  • the second image management module 301 can be configured to receive the individual tags of the user sent by the terminal 10, and instruct the second image calculation module 302 to calculate the group tags of each user according to the individual tags of different users sent by the different terminals 10.
  • the second portrait management module 301 can also send the generated group labels of different users to the terminal 10, or can be stored in a database of the portrait server 30 (for example, a distributed database such as HBase).
  • the second portrait calculation module 302 can also include a series of algorithms or models for generating a population tag.
  • the second portrait calculation module 302 can abstract a plurality of individual labels having commonality into one group label. Therefore, as shown in FIG. 9B, the second image calculation module 302 can divide a plurality of users having commonalities in a certain aspect by clustering, combination, and feature conversion according to the above-mentioned algorithm or model according to the individual tags of the plurality of users. For a group, the group label for that group can be used as a group label for users within the group.
  • the method between the group label and the individual label of the user A may be further determined by a method such as machine learning or big data mining.
  • the degree of association is such that the subsequent terminal 10 can correct the feature value of the individual tag of the user A according to the degree of association.
  • the image server 30 receives three individual tags (P1-P3) that the terminal 1 determines for the user A, and three individual tags (Q1-Q3) that the terminal 2 determines for the user B, The terminal 3 is the three individual tags (W1-W3) identified by the user C. Then, the second portrait calculation module 302 can determine that the user A belongs to the group of the “post-90s” group by clustering the individual labels, and at the same time, performing clustering by combining the characteristics of the individual labels. It is determined that the user A belongs to the group of the "American drama" group tag, and by performing feature conversion on these individual tags and then performing clustering, it is determined that the user A belongs to the group of the "game” group tag.
  • the second portrait calculation module 302 obtains the three group labels of the user A as “90” (S1), “American TV” (S2), and “game” (S3). At this time, the second portrait calculation module 302 can continue to perform big data statistics or data mining on the group label of the user A, and calculate the degree of association between each group label of the user A and each individual label. For example, the degree of association between the group tag "post-90" (S1) and the individual tag "food” (P1) is 90 points (in a case of a full score of 100), indicating that when user A is "post-90", it has The probability of the feature of "food” is about 90%. Then, the subsequent terminal 10 can correct the characteristic value of the individual label "food” (P1) originally generated by the terminal according to the degree of association.
  • the second portrait calculation module 302 can determine the group label of each user, that is, the group attribute of the user, based on the individual labels of the plurality of users, so that the terminal 10 can generate the individual label of the user and obtain the user. Group labels to generate a more complete and accurate picture of the user.
  • the second image calculation module 302 can also calculate the degree of association between the user's group tag and the individual tag, so that the terminal 10 can subsequently calibrate the feature values of the generated individual tags to further improve the final generated user image. Accuracy, thereby improving the accuracy and intelligence of the terminal 10 when providing intelligent services.
  • FIG. 11 is a schematic diagram of interaction of a method for generating a user portrait according to an embodiment of the present invention. This method is applied to the portrait system composed of the above-described terminal 10 and portrait server 30. As shown in FIG. 11, the method includes:
  • the terminal collects behavior data generated when the target user uses the terminal.
  • the target user may be collected by the data collection module 202 through one or more of system monitoring, reading a specific data interface, calling system service, and collecting a collection.
  • behavior data generated when the terminal is used for example, the behavior data may specifically include application level data, system level data, and sensor level data.
  • different acquisition periods can be set for different types of behavior data terminals.
  • the terminal may set a smaller collection period to collect user behavior data.
  • the terminal can collect the location information of the terminal, the working state of the Bluetooth, and the like every 5 minutes.
  • the terminal can set a larger acquisition period to collect user behavior data.
  • the terminal can collect the name and number of applications installed in the terminal every 24 small clocks.
  • the data collection module 202 may store the collected behavior data in a database (for example, SQLite) of the terminal, for example, store the correspondence between the collection time and the behavior data corresponding to the collection time in the form of a list in the terminal. In the database.
  • the terminal may further encrypt the collected behavior data by using an encryption algorithm (for example, AES256).
  • S1002 The terminal generates an individual label for the target user according to the foregoing behavior data, where the individual label reflects an individual behavior characteristic of the target user.
  • the first image management module 201 in the terminal may input the behavior data collected in a certain period of time into the first image calculation module 203, and the first image calculation module 203 follows the first image calculation module 203.
  • the pre-stored algorithm or model determines an individual tag that reflects the behavior characteristics of the user A based on the collected behavior data by means of machine learning or statistical analysis.
  • the behavior data sent by the first portrait management module 201 to the first portrait calculation module 203 is: the number of photographs collected in the last 24 hours. Then, when the number of photographs is greater than the first preset value (for example, 15 sheets), the first portrait calculation module 203 may determine “love photography” as one of the user labels of the user, and the corresponding feature value is 60 points ( The maximum image is taken as an example. When the number of photographs is greater than the second preset value (for example, 25 sheets, and the second preset value is greater than the first preset value), the first image calculation module 203 may determine “love photography” as One of the user's user labels. The corresponding feature value is 80 points.
  • the first preset value for example, 15 sheets
  • the first portrait calculation module 203 may determine “love photography” as one of the user labels of the user, and the corresponding feature value is 60 points ( The maximum image is taken as an example.
  • the first image calculation module 203 may determine “love photography” as One of the user's user labels.
  • the first portrait management module 201 can also generate individual tags of the target user using other algorithms or models, for example, sorting, weighting and averaging, logistic regression algorithm, Adaboost algorithm, naive Bayes algorithm, and neural network algorithm, etc., this application The embodiment does not impose any restrictions on this.
  • the personal label determined by the first image calculation module 203 for the target user may include one or more, and the embodiment of the present application does not impose any limitation on this.
  • the terminal sends the individual label of the individual label with a sensitivity less than a threshold to the image server.
  • the terminal may desensitize the individual tag (such as the address of the user A, the phone, etc.) related to the privacy of the target user generated in step S1002, so that as few individuals involved in the generated individual tags have privacy related to the user. label.
  • the individual tag such as the address of the user A, the phone, etc.
  • the terminal can determine the degree of correlation between each individual tag and the user's privacy, for example, by calculating the confidence or correlation coefficient between the individual tag and the user's privacy, etc., to obtain the sensitivity of the individual tag. .
  • the terminal may send the individual tag of the one or more target users whose sensitivity is less than the threshold to the portrait server, thereby reducing The risk of revealing user privacy when the terminal interacts with the portrait server.
  • the image server acquires an individual label of each of the N users, where N>1.
  • the N users include the target users described in steps S1001-S1003.
  • the terminal can transmit the individual tags generated for each user to the image server by performing the above steps S1001-S1003. Therefore, each time the image server receives an individual tag sent to the terminal by the terminal, It is stored in a database of the image server, thereby obtaining an individual tag for each of the N users.
  • the image server generates a group label of the target user according to the individual label of each of the N users, and the group label reflects the behavior characteristics of the group to which the target user belongs.
  • the second portrait management module 301 of the portrait server may input the individual labels of the N users to the second portrait calculation module 302, and the second portrait calculation module 302 follows the preset.
  • the algorithm or model determines the group label of each of the N users by clustering, feature combination, and feature conversion.
  • clustering refers to aggregating users with similar individual tags into a group of groups.
  • the portrait server pre-sets the correspondence between the group tag "post-90" and the group 1, which refers to a user who includes individual tags such as "love photography” and "online shopping”. Then, when the second portrait management module 301 detects that both the user A and the user B have individual tags of "love photography” and “online shopping", the user A and the user B can be regarded as two members in the group 1. Since the group tag of the group 1 is "post-90s", the group tags of the user A and the user B belonging to the group 1 also include "post-90s”.
  • Feature combination refers to converting a large number of individual tags into a small number of feature tags according to certain rules.
  • the portrait server can use these feature tags to further aggregate similar users into a group of groups by the above clustering algorithm. For example, if the user A has a total of 50 individual tags, the second portrait management module 301 can combine the 50 individual tags into four feature tags according to the four types of clothing, food, housing, and travel. Subsequently, the second portrait management module 301 can cluster the four types of four types of labels, namely, the user A, the user B, and the user C, respectively, to obtain the group label of each user. .
  • Feature conversion refers to converting a plurality of individual tags of a user into corresponding conversion tags respectively.
  • the individual tags of user A are “QQ long online”, “high transfer frequency” and “ticket”, then “QQ” Convert from long-term online to “Language”, convert “high transfer frequency” to "high income”, and convert “ticket” to "go travel”.
  • the second portrait management module 301 can use three conversion tags of "web chat”, “high income”, and “travel” to cluster with other user feature converted conversion tags, thereby obtaining a group tag of each user.
  • the image server can use more algorithms such as logistic regression algorithm, Adaboost algorithm, protocol mapping algorithm, regression analysis algorithm, Web data mining algorithm, Random Forests algorithm and K-nearest neighbors.
  • the user is assigned to a group of different characteristics, and thus the corresponding group label is given to different users, and the embodiment of the present application does not impose any limitation on this.
  • the image server can comprehensively analyze the group to which the target user belongs based on the individual tags of the plurality of users, thereby obtaining the group tag of the target user, and the group tag can reflect the behavior characteristics of the group to which the user belongs, such that the terminal subsequently combines
  • the target user's individual tags and group tags can get a more complete and accurate user portrait.
  • the group label determined by the second portrait management module 301 for the target user may include one or more, and the embodiment of the present application does not impose any limitation on this.
  • the image server generates a group tag of the target user according to the individual tags of each of the N users, which may be a continuous loop process.
  • the new user's personal tag can be used as a new input, and the individual tags of the user A-user C in FIG. 10 are used to re-determine the group to which each user belongs and The group label for each user. That is to say, the group tag determined by the image server for the user A may be a continuously updated process, and the image server may send the group tag of the user A after each update to the terminal of the user A.
  • the portrait server determines the degree of association between the target user's group tag and the individual tag.
  • the image server may further perform association rule mining under the big data on the group label to determine the group label of the target user and the individual of the target user.
  • the degree of association between tags may be performed by the image server.
  • the degree of association between the group label "post-90" and the individual label “day and night” is 90 (for example, out of 100), which means that when the target user has the "post-90" group label, there is a probability of about 90%.
  • the individual label "Day and Night” will appear.
  • the subsequent terminal can optimize the feature values of the individual tags and the individual tags generated in step S1002 according to the degree of association determined by the image server, thereby improving the accuracy of the finally generated user portrait.
  • the image server sends the group label of the target user to the terminal.
  • the association degree may be simultaneously transmitted to the terminal in step S1007.
  • the terminal corrects the individual tag generated in step S1002 according to the group tag of the target user and the degree of association, and obtains a user image of the target user.
  • the image server transmits the group tag of the target user obtained in steps S1005 and S1006 and the degree of association to the terminal, so that the terminal can generate the individual based on the behavior data in step S1002 according to the group tag and the degree of association.
  • the tag is calibrated to generate a final user portrait for the target user.
  • the individual tag generated by the terminal for the user when performing step S1002 includes two parts, one part is an individual tag with high degree of association with the user privacy, and the individual tags are not sent to the portrait server; the other part is related to the user privacy. Lower individual tags. After these individual tags are sent to the portrait server, the portrait server generates a group tag for the target user.
  • the terminal may set the group tag of the target user. And the complete set of individual tags originally generated as the user image of the target user.
  • the terminal may also input the group tag of the target user as the new behavior data into the image optimization module 204, and the image optimization module 204 recalculates the individual tags of the target user in combination with the behavior data and the group tag of the target user.
  • the individual tags generated in this way take into account the individual behavior characteristics of the target user and the group behavior characteristics of the target user, so that the optimized individual tags are more complete and accurate.
  • the image optimization module 204 may correct the individual tag generated in step S1002 according to the degree of association.
  • the terminal may add the "online shopping" tag to the individual tag of the target user, and set the "online shopping" individual tag An eigenvalue, which may be any value less than 75 minutes.
  • the degree of association between the group tag "post-90" sent by the image server and the individual tag “day and night” is 95, indicating that when the target user has the group tag "90", the probability of having the individual tag “day and night” is about It is 95%, and the characteristic value of the terminal tag "day and night” in the step S1002 is 65 points, indicating that the feature value of the feature of "day and night” judged by the terminal may be deviated. Therefore, the terminal can correct the feature value of the individual tag "day and night” based on the above-mentioned degree of association (95) on the basis of the original feature value of 65 points.
  • the corrected feature value original feature value + relevance degree * correction factor.
  • the terminal can use the corrected individual tag and the feature value of the individual tag as the user image of the target user, and the user image obtained at this time not only comprehensively considers the influence of the user's individual behavior characteristics and group behavior characteristics, but also on the user portrait.
  • the feature values of the individual tags are corrected such that the integrity and accuracy of the user image of the target user ultimately generated by the terminal is improved.
  • the terminal When the terminal receives the request for acquiring the user image by the first application, the terminal provides the user image to the first application.
  • the terminal After the terminal generates a user image with higher accuracy and more completeness for the target user, if it is detected that the application running on the terminal (for example, the first application) needs to provide the smart service to the user, the first application is used.
  • the first image management module 201 may be requested to provide a user image to the first application by calling a specific interface such as a Provider in the image query module 205. At this time, the first image management module 201 may use the user image generated in step S1008 as the result of the request. Feedback to the first application.
  • the first application can provide the user with a smarter and more convenient intelligent service.
  • the steps of performing the terminal in the above steps S1001-S1003 and S1008-S1009 may be implemented by the processor of the terminal shown in FIG. 1 executing the program instructions stored in the memory.
  • the execution steps of the image server in the above steps S1004-S1007 can be implemented by the processor of the image server executing the program instructions stored in the memory.
  • the above terminals and the like include hardware structures and/or software modules corresponding to the respective functions in order to implement the above functions.
  • Those skilled in the art will readily appreciate that the embodiments of the present application can be implemented in a combination of hardware or hardware and computer software in combination with the elements and algorithm steps of the various examples described in the embodiments disclosed herein. Whether a function is implemented in hardware or computer software to drive hardware depends on the specific application and design constraints of the solution. A person skilled in the art can use different methods to implement the described functions for each particular application, but such implementation should not be considered to be beyond the scope of the embodiments of the present application.
  • the embodiment of the present application may perform the division of the function modules on the terminal or the like according to the foregoing method example.
  • each function module may be divided according to each function, or two or more functions may be integrated into one processing module.
  • the above integrated modules can be implemented in the form of hardware or in the form of software functional modules. It should be noted that the division of the module in the embodiment of the present application is schematic, and is only a logical function division, and the actual implementation may have another division manner.
  • FIG. 3 is a schematic diagram of a possible structure of the terminal involved in the foregoing embodiment, including: a first portrait management module 201, a data collection module 202, and a first The image calculation module 203, the first image query module 205, and the image optimization module 204.
  • the related actions of these functional modules can be referred to the related description in FIG. 3, and details are not described herein again.
  • FIG. 8 is a schematic diagram showing a possible configuration of the image server involved in the above embodiment, including: a second image management module 301 and a second image calculation module 302. .
  • the related actions of the functional modules may be referred to in the related description of FIG. 8 and will not be further described herein.
  • FIG. 13 a possible structural diagram of the terminal involved in the above embodiment is shown, including a processing module 2101, a communication module 2102, an input/output module 2103, and a storage. Module 2104.
  • the processing module 2101 is configured to control and manage the action of the terminal.
  • the communication module 2102 is configured to support communication between the terminal and other network entities.
  • the input/output module 2103 is for receiving information input by a user or outputting information provided to the user and various menus of the terminal.
  • the storage module 2104 is configured to save program codes and data of the terminal.
  • FIG. 14 a possible schematic diagram of the image server involved in the above embodiment is shown, including a processing module 2201, a communication module 2202, and a storage module 2203.
  • the processing module 2201 is configured to control and manage the action of the image server.
  • the communication module 2202 is configured to support communication between the portrait server and other servers or terminals.
  • the storage module 2203 is configured to save program code and data of the image server.
  • the processing module 210 1/2201 may be a processor or a controller, and may be, for example, a central processing unit (CPU), a GPU, a general-purpose processor, and a digital signal processor (DSP).
  • DSP digital signal processor
  • ASIC Application-Specific Integrated Circuit
  • FPGA Field Programmable Gate Array
  • the processor may also be a combination of computing functions, for example, including one or more microprocessor combinations, a combination of a DSP and a microprocessor, and the like.
  • the above communication module 2102/2202 may be a transceiver, a transceiver circuit, or a communication interface or the like.
  • the communication module 1303 may specifically be a Bluetooth device, a Wi-Fi device, a peripheral interface, or the like.
  • the above-described input/output module 2103 may be a touch screen, a display, a microphone, or the like that receives information input by a user or outputs information provided to a user.
  • the display may be configured in the form of a liquid crystal display, an organic light emitting diode or the like.
  • a touch panel can be integrated on the display for collecting touch events on or near the display, and transmitting the collected touch information to other devices (such as a processor, etc.).
  • the above memory modules 2104/2203 may be memories, which may include high speed random access memories (RAM), and may also include nonvolatile memories such as magnetic disk storage devices, flash memory devices, or other volatile solid state storage devices.
  • RAM high speed random access memories
  • nonvolatile memories such as magnetic disk storage devices, flash memory devices, or other volatile solid state storage devices.
  • the computer program product includes one or more computer instructions.
  • the computer can be a general purpose computer, a special purpose computer, a computer network, or other programmable device.
  • the computer instructions can be stored in a computer readable storage medium or transferred from one computer readable storage medium to another computer readable storage medium, for example, the computer instructions can be from a website site, computer, server or data center Transfer to another website site, computer, server, or data center by wire (eg, coaxial cable, fiber optic, digital subscriber line (DSL), or wireless (eg, infrared, wireless, microwave, etc.).
  • the computer readable storage medium can be any available media that can be accessed by a computer or a data storage device such as a server, data center, or the like that includes one or more available media.
  • the usable medium may be a magnetic medium (eg, a floppy disk, a hard disk, a magnetic tape), an optical medium (eg, a DVD), or a semiconductor medium (such as a solid state disk (SSD)).

Landscapes

  • Business, Economics & Management (AREA)
  • Strategic Management (AREA)
  • Engineering & Computer Science (AREA)
  • Accounting & Taxation (AREA)
  • Development Economics (AREA)
  • Finance (AREA)
  • Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Marketing (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Information Transfer Between Computers (AREA)

Abstract

本申请的实施例提供一种用户画像的生成方法及装置，涉及智慧化技术领域，可提高终端生成的用户画像的准确度。该方法包括：终端将为用户生成的至少一个个体标签发送至画像服务器，所述个体标签反映了所述用户的个人行为特征；所述终端接收所述画像服务器为所述用户生成的至少一个群体标签，所述群体标签是所述画像服务器至少基于所述至少一个个体标签生成的，所述群体标签反映了所述用户所属群体的行为特征；所述终端使用所述群体标签更新所述用户的用户画像；所述终端向第一应用提供更新后的所述用户画像中的至少一部分。

Description

一种用户画像的生成方法及装置 技术领域
本申请实施例涉及智慧化技术领域，尤其涉及一种用户画像的生成方法及装置。
背景技术
随着信息通信技术(information communication technology，ICT)的不断发展，物理世界中的人类活动越来越多的深入到数字世界中。
在数字世界中，手机等终端可以根据用户的使用行为将实际用户抽象为具有一个或多个标签的用户画像。例如，用户A经常使用手机在晚上12点后看动漫，那么，手机可将“晚睡”、“二次元”等标签作为用户A的用户画像。后续，手机可基于用户A的用户画像为用户提供定制化的服务和功能，以提高手机的工作效率。
一个用户完整的用户画像中通常包含多个标签，这些标签中有一部分是直接基于用户在手机上的个体使用行为生成的个体标签，例如，上述“晚睡”的标签是基于用户A玩手机的时间生成的。但上述标签中还有一部分标签是需要通过对不同用户的使用行为进行大数据计算和数据挖掘才能得到的群体标签，例如，服务器可以通过对比多个用户的用户画像确定用户A属于“90后”这一群体标签。
那么，如果由服务器为用户生成用户画像，则需要终端将用户的行为数据上传给服务器，但很多涉及用户隐私的行为数据无法上传到服务器，导致服务器生成的用户画像的准确度下降。相应的，如果由终端为用户生成用户画像，由于终端仅能采集到使用该终端的单个用户的行为数据，因此终端无法确定出该用户所属的群体标签，也会导致生成的用户画像的准确度下降。
发明内容
本申请的实施例提供一种用户画像的生成方法及装置，可提高终端生成的用户画像的准确度。
为达到上述目的，本申请的实施例采用如下技术方案：
第一方面，本申请的实施例提供一种用户画像的生成方法，包括：终端将为用户生成的至少一个个体标签发送至画像服务器，该个体标签反映了该用户的个人行为特征；终端接收画像服务器为该用户生成的至少一个群体标签(该群体标签反映了该用户所属群体的行为特征)，该群体标签是画像服务器至少基于该至少一个个体标签生成的；这样，终端可以使用该群体标签更新用户的用户画像，从而向第一应用提供更新后的该用户画像中的至少一部分。
此时，更新后的该用户画像中不仅反映了用户的个体行为特征，同时反映了用户所属群体的群体行为特征，因此该更新后的该用户画像的完整性和准确率较高，使得第一应用使用该更新后的该用户画像提供的业服服务更加智能和准确。
在一种可能的设计方法中，在终端将为用户生成的至少一个个体标签发送至画像服务器之前，还包括：终端采集该用户使用终端时产生的行为数据；终端根据该行为数据为该用户生成至少一个个体标签，每个个体标签中包括该个体标签的类型和该个体标签的特征值。
在一种可能的设计方法中，终端将为用户生成的至少一个个体标签发送至画像服务器，包括：终端将该至少一个个体标签中敏感度小于阈值的个体标签发送至画像服务器，该敏感度用于指示该个体标签与用户隐私之间的相关程度，从而在提高用户画像准确率的同时降低用户隐私泄露的风险。
在一种可能的设计方法中，终端使用该群体标签更新该用户的用户画像，包括：终端将该群体标签添加至该用户的用户画像中，得到更新后的用户画像，该更新后的用户画像中包括该群体标签和终端生成的至少一个个体标签。
在一种可能的设计方法中，终端使用该群体标签更新该用户的用户画像，包括：终端根据该行为数据和该群体标签更新终端生成的至少一个个体标签，得到更新后的用户画像，该更新后的用户画像中包括该更新后的个体标签。
在一种可能的设计方法中，上述更新后的用户画像中还包括该群体标签。
在一种可能的设计方法中，在终端接收画像服务器为该用户生成的至少一个群体标签之后，还包括：终端接收画像服务器为该用户生成的第一群体标签与第一个体标签之间的关联度，第一群体标签为该至少一个群体标签中的一个，第一个体标签为该至少一个个体标签中的一个；终端根据该关联度校正第一个体标签的特征值，从而进一步提高后续终端生成的用户画像的准确度。
示例性的，终端根据该关联度校正第一个体标签的特征值，具体包括：终端将第一个体标签的特征值与校正值之和作为第一个体标签校正后的特征值，该校正值为该关联度与预设的校正因子的乘积，该校正因子用于反映该关联度对第一个体标签的影响程度。
第二方面，本申请的实施例提供一种用户画像的生成方法，包括：画像服务器获取至少一个用户的个体标签；画像服务器根据该至少一个用户中每个用户的个体标签，生成目标用户的群体标签，该群体标签反映了该目标用户所属群体的行为特征，该目标用户为该至少一个用户中的一个；画像服务器将该目标用户的群体标签发送给终端。
在一种可能的设计方法中，画像服务器根据该至少一个用户中每个用户的个体标签，生成目标用户的群体标签，包括：画像服务器根据该至少一个用户中每个用户的个体标签，将该至少一个用户划分为至少一个群体；画像服务器将该目标用户所属群体的标签作为该目标用户的群体标签。
在一种可能的设计方法中，画像服务器根据该至少一个用户中每个用户的个体标签，将该至少一个用户划分为至少一个群体，包括：画像服务器基于该至少一个用户中每个用户的个体标签，通过聚类、特征组合后聚类以及特征转换后聚类中的一种或多种方式，将该至少一个用户划分为至少一个群体。
在一种可能的设计方法中，在画像服务器获取至少一个用户的个体标签之后，还包括：画像服务器确定该目标用户的群体标签分别与该目标用户的每个个体标签之间的关联度；画像服务器将该关联度发送给终端。
第三方面，本申请的实施例提供一种终端，包括：包括画像管理模块，以及与该画像管理模块均相连的数据采集模块、画像计算模块、画像优化模块以及画像查询模块，其中，该画像管理模块，用于：将为用户生成的至少一个个体标签发送至画像服务器，该个体标签反映了该用户的个人行为特征；接收画像服务器为该用户生成的至 少一个群体标签，该群体标签是画像服务器至少基于该至少一个个体标签生成的，该群体标签反映了该用户所属群体的行为特征；该画像优化模块，用于：使用该群体标签更新该用户的用户画像；该画像查询模块，用于：向第一应用提供更新后的该用户画像中的至少一部分。
在一种可能的设计方法中，该数据采集模块，用于：采集该用户使用终端时产生的行为数据；该画像计算模块，用于：根据该行为数据为该用户生成至少一个个体标签，每个个体标签中包括该个体标签的类型和该个体标签的特征值。
在一种可能的设计方法中，该画像管理模块，具体用于：将该至少一个个体标签中敏感度小于阈值的个体标签发送至画像服务器，该敏感度用于指示该个体标签与用户隐私之间的相关程度。
在一种可能的设计方法中，该画像优化模块，具体用于：将该群体标签添加至该用户的用户画像中，得到更新后的用户画像，该更新后的用户画像中包括该群体标签和终端生成的至少一个个体标签。
在一种可能的设计方法中，该画像优化模块，具体用于：根据该行为数据和该群体标签更新终端生成的至少一个个体标签，得到更新后的用户画像，该更新后的用户画像中包括该更新后的个体标签。
在一种可能的设计方法中，该更新后的用户画像中还包括该群体标签。
在一种可能的设计方法中，该画像管理模块，还用于：接收画像服务器为该用户生成的第一群体标签与第一个体标签之间的关联度，第一群体标签为该至少一个群体标签中的一个，第一个体标签为该至少一个个体标签中的一个；该画像优化模块，还用于：根据该关联度校正第一个体标签的特征值。
在一种可能的设计方法中，该画像优化模块，具体用于：将第一个体标签的特征值与校正值之和作为第一个体标签校正后的特征值，该校正值为该关联度与预设的校正因子的乘积，该校正因子用于反映该关联度对第一个体标签的影响程度。
第四方面，本申请的实施例提供一种服务器，包括画像管理模块，以及与该画像管理模块相连的画像计算模块，其中，该画像管理模块，用于：获取至少一个用户的个体标签；该画像计算模块，用于：根据该至少一个用户中每个用户的个体标签，生成目标用户的群体标签，该群体标签反映了该目标用户所属群体的行为特征，该目标用户为该至少一个用户中的一个；该画像管理模块，还用于：将该目标用户的群体标签发送给终端。
在一种可能的设计方法中，该画像计算模块，具体用于：根据该至少一个用户中每个用户的个体标签，将该至少一个用户划分为至少一个群体；将该目标用户所属群体的标签作为该目标用户的群体标签。
在一种可能的设计方法中，该画像计算模块，具体用于：基于该至少一个用户中每个用户的个体标签，通过聚类、特征组合后聚类以及特征转换后聚类中的一种或多种方式，将该至少一个用户划分为至少一个群体。
在一种可能的设计方法中，该画像计算模块，还用于：确定该目标用户的群体标签分别与该目标用户的每个个体标签之间的关联度；该画像管理模块，还用于：将该关联度发送给终端。
第五方面，本申请的实施例提供一种终端，包括：处理器、存储器、总线和通信接口；该存储器用于存储计算机执行指令，该处理器与该存储器通过该总线连接，当终端运行时，该处理器执行该存储器存储的该计算机执行指令，以使终端执行上述任一项用户画像的生成方法。
第六方面，本申请的实施例提供一种画像服务器，包括：处理器、存储器、总线和通信接口；该存储器用于存储计算机执行指令，该处理器与该存储器通过该总线连接，当画像服务器运行时，该处理器执行该存储器存储的该计算机执行指令，以使画像服务器执行上述任一项用户画像的生成方法。
第七方面，本申请实施例提供一种计算机可读存储介质，该计算机可读存储介质中存储有指令，当该指令在上述任一项终端上运行时，使得终端执行上述任一项用户画像的生成方法。
第八方面，本申请实施例提供一种计算机可读存储介质，该计算机可读存储介质中存储有指令，当该指令在上述任一项画像服务器上运行时，使得画像服务器执行上述任一项用户画像的生成方法。
第九方面，本申请实施例提供一种包含指令的计算机程序产品，当其在上述任一项终端上运行时，使得终端执行上述任一项用户画像的生成方法。
第十方面，本申请实施例提供一种包含指令的计算机程序产品，当其在上述任一项画像服务器上运行时，使得画像服务器执行上述任一项用户画像的生成方法。
本申请的实施例中，上述终端或画像服务器中各部件的名字对设备本身不构成限定，在实际实现中，这些部件可以以其他名称出现。只要各个部件的功能和本申请的实施例类似，即属于本申请权利要求及其等同技术的范围之内。
另外，第二方面至第十方面中任一种设计方式所带来的技术效果可参见上述第一方面中不同设计方法所带来的技术效果，此处不再赘述。
附图说明
图1为本申请实施例提供的一种终端的结构示意图一；
图2为本申请实施例提供的一种用户画像平台的结构示意图；
图3为本申请实施例提供的一种用户画像模块的结构示意图一；
图4为本申请实施例提供的一种行为数据的示意图；
图5为本申请实施例提供的一种用户画像模块的结构示意图二；
图6为本申请实施例提供的一种用户标签的示意图；
图7为本申请实施例提供的一种用户画像模块的结构示意图三；
图8为本申请实施例提供的一种画像服务器的结构示意图；
图9A为本申请实施例提供的一种群体标签的生成原理示意图一；
图9B为本申请实施例提供的一种群体标签的生成原理示意图二；
图10为本申请实施例提供的一种群体标签的生成原理示意图三；
图11为本申请实施例提供的一种用户画像的生成方法的流程示意图；
图12为本申请实施例提供的一种用户画像的生成方法的原理示意图；
图13为本申请实施例提供的一种终端的结构示意图二；
图14为本申请实施例提供的一种画像服务器的结构示意图。
具体实施方式
随着智慧化业务的发展，可以基于用户的历史行为习惯或者基于一些规则或模型，在终端上进行一些智能的提醒或者服务，以便于用户更方便的使用终端，使得用户觉得终端越来越智能化。
终端可以通过自身或者通过与云端的结合，来实现各种智慧化业务。具体的，终端可以包括规则平台、算法平台和用户画像模块。终端可以通过这三个平台中的一个或多个以及其它资源实现各种智慧化业务，例如：1、服务推荐业务；2、提醒业务；3、通知过滤业务。
1、服务推荐业务。
终端包括用于实现该服务推荐业务的推荐服务框架(framework)，该推荐服务框架至少可以包括算法平台、规则平台和用户画像模块。
上述规则平台可以根据规则匹配出该终端的用户在当前场景下希望使用的服务。上述算法平台可以根据模型预测出该终端的用户在当前场景下希望使用的服务。该推荐服务框架可以将所述规则平台或所述算法平台预测出的服务置于推荐应用的显示界面中，以便于用户可以方便的通过该推荐应用的显示界面进入该服务对应的界面。
其中，上述规则可以由服务器(即云端)下发给终端。该规则可以通过大数据统计获取，也可以根据经验数据归纳得到。上述模型可以通过以下方式获取：通过所述算法平台训练用户历史数据和用户特征数据，得到模型。并且基于新的用户数据和特征数据可以更新该模型。
其中，用户历史数据可以为用户在一段时间段内使用该终端的行为数据。用户特征数据可以包括用户画像(user profile)或其他类型的特征数据，所述其他类型的特征数据例如可以为当前用户的行为数据。其中，用户画像可以通过终端中的所述用户画像模块得到。
2、提醒业务。
终端包括用于实现该提醒业务的推荐框架(framework)。该推荐框架至少可以包括规则平台、图形用户界面(graphical user interface)和用户画像模块。
上述规则平台可以监听各种事件。终端中的应用可以向该规则平台注册各种规则；然后该规则平台根据注册的规则，监听终端中的各种事件；将监听到的事件与规则进行匹配，并当监听到的事件与某个规则的所有条件都匹配时，触发该规则对应的提醒，即向用户推荐一个亮点事件。最终由图形用户界面显示该提醒或者由注册规则的应用显示该提醒。其中，一些规则的条件可以为对用户画像的限定。该规则平台可以向该用户画像模块请求当前的用户画像，以判断当前的用户画像是否与规则中的条件相匹配。
3、通知过滤业务。
终端包括用于实现该通知过滤业务的通知过滤框架(framework)。该通知过滤框架至少可以包括规则平台、算法平台和用户画像模块。
上述通知过滤框架在获取到一个通知时，可以通过该规则平台确定该通知的类型，也可以通过该算法平台确定该通知的类型。然后根据该通知的类型以及用户的喜好，确定该通知是否为用户感兴趣的通知，并且对于用户感兴趣的通知和用户不感兴趣的 通知，进行不同方式的提醒显示。用户的喜好可以包括用户画像，也可以包括用户对某类通知的历史处理行为。其中，用户画像是由该用户画像模块提供的。
需要说明的是，终端可以包括一个规则平台，该规则平台向上述三种框架提供每个框架所需的能力。终端也可以包括多个规则平台，这多个规则平台分别向上述三种框架提供能力。同样的，终端可以包括一个算法平台，该算法平台向上述推荐服务框架和通知过滤框架提供每个框架所需的能力；或者，终端也可以包括两个算法平台，分别向这两个框架提供能力。终端可以包括一个用户画像模块，该用户画像模块向上述三种框架提供每个框架所需的能力。或者，终端也可以包括多个用户画像模块，分别向每个框架提供能力。
本申请以下各实施例主要对上述的用户画像模块进行详细介绍。
本发明实施例提供的用户画像模块可以包含在终端中。该终端例如可以为：移动电话、平板电脑(tablet personal computer)、膝上型电脑(laptop computer)、数码相机、个人数字助理(personal digital assistant，PDA)、导航装置、移动上网装置(mobile internet device，MID)或可穿戴式设备(wearable device)等。
图1为本发明实施例提供的终端的部分结构框图。该终端以手机100为例进行说明，参考图1，手机100包括：射频(radio frequency，RF)电路110、电源120、处理器130、存储器140、输入单元150、显示单元160、传感器170、音频电路180、以及无线保真(wireless-fidelity，Wi-Fi)模块190等部件。本领域技术人员可以理解，图1中示出的手机结构并不构成对手机的限定，可以包括比图示更多或更少的部件，或者组合某些部件，或者不同的部件布置。
下面结合图1对手机100的各个构成部件进行具体的介绍：
RF电路110可用于收发信息或在通话过程中进行信号的接收和发送。例如：RF电路110可以将从基站接收的下行数据发送给处理器130处理，并把上行数据发送给基站。
通常，RF电路包括但不限于RF芯片、天线、至少一个放大器、收发信机、耦合器、低噪声放大器(low noise amplifier，LNA)、双工器、射频开关等。此外，RF电路110还可以与网络和其他设备进行无线通信。所述无线通信可以使用任一通信标准或协议，包括但不限于全球移动通讯***(global system of mobile communication，GSM)、通用分组无线服务(general packet radio service，GPRS)、码分多址(code division multiple access，CDMA)、宽带码分多址(wideband code division multiple access，WCDMA)、长期演进(long term evolution，LTE)、电子邮件、短消息服务(short messaging service，SMS)等。
存储器140可用于存储软件程序以及模块，处理器130通过运行存储在存储器140的软件程序以及模块，从而执行手机100的各种功能应用以及数据处理。存储器140可主要包括存储程序区和存储数据区，其中，存储程序区可存储操作***、至少一个功能所需的应用程序(比如声音播放功能、图像播放功能等)等；存储数据区可存储根据手机100的使用所创建的数据(比如音频数据、电话本等)等。此外，存储器140可以包括高速随机存取存储器，还可以包括非易失性存储器，例如至少一个磁盘存储 器件、闪存器件、或其他易失性固态存储器件。存储器140还可以存储知识库、标签库和算法库。
输入单元150可用于接收输入的数字或字符信息，以及产生与手机100的用户设置以及功能控制有关的键信号输入。具体地，输入单元150可包括触控面板151以及其他输入设备152。触控面板151，也称为触摸屏，可收集用户在其上或附近的触摸操作(比如用户使用手指、触笔等任何适合的物体或附件在触控面板151上或在触控面板151附近的操作)，并根据预先设定的程式驱动相应的连接装置。可选的，触控面板151可包括触摸检测装置和触摸控制器两个部分。其中，触摸检测装置检测用户的触摸方位，并检测触摸操作带来的信号，将信号传送给触摸控制器；触摸控制器从触摸检测装置上接收触摸信息，并将它转换成触点坐标，再送给处理器130，并能接收处理器130发来的命令并加以执行。此外，可以采用电阻式、电容式、红外线以及表面声波等多种类型实现触控面板151。除了触控面板151，输入单元150还可以包括其他输入设备152。具体地，其他输入设备152可以包括但不限于物理键盘、功能键(比如音量控制按键、开关按键等)、轨迹球、鼠标、操作杆等中的一种或多种。
显示单元160可用于显示由用户输入的信息或提供给用户的信息以及手机100的各种菜单。显示单元160可包括显示面板161，可选的，可以采用液晶显示屏(liquid crystal display，LCD)、机电激光显示(organic light-emitting diode，OLED)等形式来配置显示面板161。进一步的，触控面板151可覆盖显示面板161，当触控面板151检测到在其上或附近的触摸操作后，传送给处理器130以确定触摸事件的类型，随后处理器130根据触摸事件的类型在显示面板161上提供相应的视觉输出。虽然在图1中，触控面板151与显示面板161是作为两个独立的部件来实现手机100的输入和输入功能，但是在某些实施例中，可以将触控面板151与显示面板161集成而实现手机100的输入和输出功能。
手机100还可包括至少一种传感器170，比如光传感器、运动传感器以及其他传感器。具体地，光传感器可包括环境光传感器及接近传感器，其中，环境光传感器可根据环境光线的明暗来调节显示面板161的亮度，接近传感器可在手机100移动到耳边时，关闭显示面板161和/或背光。作为运动传感器的一种，加速计传感器可检测各个方向上(一般为三轴)加速度的大小，静止时可检测出重力的大小及方向，可用于识别手机姿态的应用(比如横竖屏切换、相关游戏、磁力计姿态校准)、振动识别相关功能(比如计步器、敲击)等。手机100还可以配置的陀螺仪、气压计、湿度计、温度计、红外线传感器等其他传感器，在此不再赘述。
音频电路180、扬声器181、麦克风182可提供用户与手机100之间的音频接口。音频电路180可将接收到的音频数据转换后的电信号，传输到扬声器181，由扬声器181转换为声音信号输出；另一方面，麦克风182将收集的声音信号转换为电信号，由音频电路180接收后转换为音频数据，再将音频数据输出至RF电路110以发送给比如另一手机，或者将音频数据输出至存储器140以便进一步处理。
Wi-Fi属于短距离无线传输技术，手机100通过Wi-Fi模块190可以帮助用户收发电子邮件、浏览网页和访问流式媒体等，它为用户提供了无线的宽带互联网访问。虽然图1示出了Wi-Fi模块190，但是可以理解的是，其并不属于手机100的必须构成， 完全可以根据需要在不改变发明的本质的范围内而省略。
处理器130是手机100的控制中心，利用各种接口和线路连接整个手机的各个部分，通过运行或执行存储在存储器140内的软件程序和/或模块，以及调用存储在存储器140内的数据，执行手机100的各种功能和处理数据，从而实现基于手机的多种业务。可选的，处理器130可包括一个或多个处理单元；优选的，处理器130可集成应用处理器和调制解调处理器，其中，应用处理器主要处理操作***、用户界面和应用程序等，调制解调处理器主要处理无线通信。可以理解的是，上述调制解调处理器也可以不集成到处理器130中。
本发明实施例中，处理器130可以执行存储器140中存储的程序指令，来在实现以下实施例所示的方法。
手机100还包括给各个部件供电的电源120(比如电池)，优选的，电源可以通过电源管理***与处理器130逻辑相连，从而通过电源管理***实现管理充电、放电、以及功耗等功能。
尽管未示出，手机100还可以包括摄像头、蓝牙模块等，在此不予赘述。
本发明实施例提供的终端中包括用户画像模块，该用户画像模块可以通过收集与分析使用该终端的用户的各种行为数据，抽象出一个用户的信息全貌。根据应用的请求，该用户画像模块通过抽象出的信息全貌可以预测用户当前可能的行为或喜好，并将预测的结果返回给应用，即向应用返回用户画像(User Profile)。
其中，用户画像中通常包括一个或多个用于反映用户特征的用户标签，一个用户标签可以划分为两部分，一部分是用户标签类型，另一部分是该用户标签的特征值。以用户A为例，如表1所示，用户A的用户画像中包括4个用户标签，其中，用户标签1的类型为“性别”，该用户标签1的特征值为女性，即说明用户A的性别为女性；用户标签2的类型为“住址”，该用户标签2的特征值为北京市，即说明用户A住在北京市；用户标签3的类型为“熬夜”，该用户标签3的特征值为“85分”(以满分为100分举例)，即说明用户A产生熬夜行为的概率较高。当用户B也具有“熬夜”类型的用户标签时，如果其打分为“60分”，则说明用户B产生熬夜行为的概率小于用户A产生熬夜行为的概率。
表1
Figure PCTCN2018073673-appb-000001
进一步地，用户画像中的用户标签可划分为个体标签和群体标签两种类型。
其中，个体标签是指可基于用户使用终端时的行为数据直接抽象出的用户特征，个体标签一般反映的是用户的个人行为特征。例如，用户经常在晚上12点之后使用手机，那么手机可以根据用户的使用习惯确定出用户具有“熬夜”类型的个体标签及该 个体标签的特征值(即打分情况)。
而群体标签是指需要通过多个用户的行为数据进行聚类等分析后，根据某一用户所属群体的特征，对该用户生成的标签。例如，根据用户A、用户B以及用户C的行为数据可以确定用户A、用户B以及用户C均属于具有“熬夜”特征的一个群体，该群体的群体标签包括“工作狂”，那么，可将“工作狂”作为用户A、用户B以及用户C的一个群体标签。
图2为本发明实施例提供的一种用户画像平台的架构示意图。如图2所示，该用户画像平台中包括至少一个终端10和画像服务器30，其中，终端10中包括用户画像模块20。
上述用户画像模块20可以为终端10中的多种应用提供用户画像。该应用可以为***级应用，也可以为普通级别应用。***级应用一般指的是：该应用具有***级权限，可以获取各种***资源。普通级别应用一般指的是：该应用具有普通权限，可能无法获取某些***资源，或者需要用户授权，才能获取一些***资源。
***级应用可以为该终端10中预装的应用。普通级别应用可以为终端10中预装的应用，也可以为后续用户自行安装的应用。例如：用户画像模块20可以分别向服务推荐应用、提醒应用、通知过滤应用等***级应用提供用户画像。其中，服务推荐应用、提醒应用、通知过滤应用分别用于实现上述实施例中的服务推荐业务、提醒业务、通知过滤业务。当然，用户画像模块20还可以为视频应用、新闻应用或其它应用提供用户画像。
用户画像模块20还可以与云侧(即网络侧)的画像服务器30进行通信。
在本申请实施例中，用户画像模块20可以将生成的用户画像中不涉及用户隐私的个体标签发送给画像服务器30，降低用户隐私泄露的风险。同时，画像服务器30接收到某一用户(例如用户A)的个体标签后，可以结合其他终端10发送的其他一个或多个用户的个体标签，通过特征聚类、组合或者特征转换等方法确定用户A所属的群体，从而生成用户A的群体标签。
后续，画像服务器30可将为用户A生成的群体标签发送给终端10，由用户画像模块20结合用户A的个体标签和用户A的群体标签为用户A生成完整性和准确率较高的用户画像，从而提高终端10使用的用户画像的准确度。
图3为本发明实施例提供的一种终端10中用户画像模块的架构示意图。如图3所示，用户画像模块20可以包括第一画像管理模块201、数据采集模块202、第一画像计算模块203、画像优化模块204、以及画像查询模块205。
数据采集模块202
数据采集模块202为用户画像模块20提供基础元数据的采集能力支持。数据采集模块202可以采集用户使用终端10时产生的行为数据，并对采集到的行为数据进行存储和读写管理。
具体的，图4为本发明实施例提供的行为数据的示意图，如图4所示，数据采集模块202采集的行为数据具体可以包括应用层级数据401、***层级数据402以及传 感器层级数据403。
其中，应用层级数据401可以包括应用层的应用在运行时收集到的可反映用户行为特征的数据，例如，应用名、应用使用时间、使用时长等。示例性的，当正在运行的应用为视频应用时，数据采集模块202还可以采集正在播放的视频名、视频停止时间、视频播放集数、视频总集数等；当正在运行的应用为音乐应用时，数据采集模块202还可以采集正在播放的音乐名、音乐类型、播放时长、播放频率等；当正在运行的应用为美食应用时，数据采集模块202还可以采集当前的店铺名、美食类型、店铺地址等。在采集用户的行为数据时，数据采集模块202还可以根据具体情况，使用图片文本感知技术来采集数据，例如：通过光学字符识别(optical character recognition，OCR)技术识别图片中的文字内容，以获取图片中的文本信息。
***层级数据402可以包括框架层(framework)中提供的各种服务在运行时收集到的可反映用户行为特征的数据。例如，数据采集模块202可以通过监听服务，监听来自操作***或应用的广播消息，获取蓝牙开关状态、SIM卡状态、应用运行状态、自动旋转开关状态、热点开关状态等信息；又例如，数据采集模块202可以通过调用特定的接口，例如安卓***提供的联系人提供接口(contact provider API)、内容提供接口(content provider API)、日历提供接口(calender provider API)等，获取***的实时场景信息，例如，音频、视频、图片、通讯录、日程安排、时间、日期、电量、网络状态、耳机状态等信息。
传感器层级数据403可以包括通过传感器等器件收集到的用于反映用户行为特征的数据。例如距离传感器、加速度传感器、气压传感器、重力传感器或陀螺仪等传感器运行时产生的数据，通过这些数据可以识别出用户处于以下的行为状态：车载、骑行、走路、跑步、静止和其他。
在本申请实施例中，可以将数据采集模块202的采集周期设置为时长较短的采集周期，例如，采集周期可以为不超过24小时的任意取值。例如，数据采集模块202可以每隔5分钟采集终端10的GPS数据，每隔24小时采集终端10内图库中存储的图像数量。这样，终端10只需维护最近24小时内采集到的用户的行为数据，避免占用终端10过多的计算资源和存储资源。
示例性的，数据采集模块202可以通过***监听、读取特定数据接口、调用***服务、打点采集等方式采集上述应用层级数据401、***层级数据402以及传感器层级数据403。
第一画像计算模块203
第一画像计算模块203中可包括一系列个体标签的生成算法或模型，第一画像计算模块203用于接收数据采集模块202在一定时间内采集到的用户的行为数据，以及按照上述算法或模型确定用户的个体标签。
具体的，如图5所示，可由第一画像管理模块201将数据采集模块202在最近24小时内采集到的行为数据发送给第一画像计算模块203，由第一画像计算模块203按照上述算法或模型通过统计分析、机器学习等方法确定出反映用户行为特征的多个个体标签。
这些个体标签中的有些标签，例如用户的住址、电话等可能涉及用户隐私，因此， 第一画像计算模块203还可以对涉及用户隐私的个体标签进行脱敏处理，降低个体标签的敏感度。
示例性的，如图6所示，用户的个体标签包括但不限于以下六类标签：基本属性、社会属性、行为习惯、兴趣爱好、心理学属性和手机使用喜好等。
其中，上述基本属性包括但不限于：个人信息和生理特征。所述个人信息包括但不限于：姓名、年龄、证件类型、学历、星座、信仰、婚姻状态和邮箱。
上述社会属性包括但不限于：行业/职业、职务、收入水平、孩子状态、车辆使用情况、房屋居住、手机和移动运营商。所述房屋居住可以包括：租房、自有房和还贷中。所述手机可以包括：品牌和价位。所述移动运营商可以包括：品牌、网络、流量特点和手机号码。所述品牌可以包括：移动、联通、电信和其它。所述网络可以包括：无、2G、3G和4G。所述流量特点可以包括：高、中和低。
上述行为习惯包括但不限于：地理位置、生活习惯、交通方式、居住酒店类型、经济/理财特性、餐饮习惯、购物特性和支付情况。所述生活习惯可以包括：作息时间、在家时间、上班时间、电脑上网时间和买菜购物时间。所述购物特性可以包括：购物品类和购物方式。所述支付情况可以包括：支付时间、支付地点、支付方式、单次支付金额和支付总金额。
上述兴趣爱好包括但不限于：读书喜好、新闻喜好、视频喜好、音乐喜好、运动喜好和旅游喜好。所述读书喜好可以包括：读书频率、读书时间段、读书总时长和读书分类。
上述心理学属性包括但不限于：生活方式、个性和价值观。
上述手机使用喜好包括但不限于：应用喜好、通知提醒、应用内操作、用户常用、***应用和常用设置。
那么，第一画像管理模块201通过统计分析、机器学习等方法确定用户的个体标签后，可结合当前用户所处的动态场景，例如，当前时间、当前位置(经纬度)、运动状态、天气、地点(POI)、手机状态和开关状态等，得到对当前实时场景的感知结果，例如，感知结果为上班路上、旅行中等。那么，基于对当前实时场景的感知结果，终端可对用户在终端上的后续行为进行预测，从而提供智能化的定制化个***，例如在用户的下班时间自动为用户显示回家路线和路况等。
需要说明的是，上述的各种个体标签仅作为举例。在具体实现方式中，第一画像计算模块203中维护中的具体个体标签可以随业务的需求进行扩展，可以增加新的类型的标签，也可以对已有的标签进行更细化的分类，第一画像计算模块203为用户生成的个体标签可反映出该用户的个性化特征。
进一步地，第一画像计算模块203生成用户的个体标签后，一方面可以将该个体标签像保存至终端10的终端的数据库(例如，SQLite)中缓存一定时间(例如7天)，另一方面，可由第一画像管理模块201将该个体标签中不涉及用户隐私的标签发送给画像服务器30。
另外，终端10可使用预设的加密算法，例如，高级加密标准(Advanced Encryption Standard，AES)对上述个体标签加密，并将加密后的个体标签存储在SQLite中，以提高个体标签在终端10内的安全性。
第一画像管理模块201
第一画像管理模块201与上述数据采集模块202、第一画像计算模块203、画像优化模块204以及画像查询模块205均耦合。
具体的，第一画像管理模块201是终端10中提供用户画像服务的控制中心，可用于提供用户画像服务的各项管理功能及运行脚本，例如，启动建立用户画像的服务、从数据采集模块202中获取用户的行为数据、指示第一画像计算模块203计算用户的个体标签、指示画像优化模块204对生成包括用户的个体标签和全体标签的完整用户画像、指示画像查询模块205对用户身份进行鉴权或者向APP提供用户画像、更新算法库、清理过期数据、与画像服务器30同步数据等。
示例性的，第一画像管理模块201获取到第一画像计算模块203为用户生成的个体标签后，可将该个体标签中不涉及用户隐私的一个或多个同步至画像服务器30。例如，终端10可基于网络协议(hypertext transfer protocol over secure socket layer，HTTPS)协议中的post/get请求方法，将生成的个体标签发送给画像服务器30。
这样，终端10向画像服务器30发送的个体标签中不会泄露用户的个人隐私，后续画像服务器30可根据接收到的个体标签为用户确定其所属的群体标签，从而得到完整、准确的用户画像。
画像优化模块204
如图7所示，第一画像管理模块201可以将第一画像计算模块203生成的个体标签，以及画像服务器30发来的用户的群体标签输入给画像优化模块204。
这样，画像优化模块204可将上述群体标签作为新增的行为数据，结合原始采集到的行为数据生成完整的用户画像，由于生成该用户画像时综合考虑了用户的个体行为特征以及用户所属群体的群体行为特征，因此画像优化模块204得到的用户画像中既包括用户的个体标签，又包括用户的群体标签，从而可以提高用户画像的完整性和准确度。
进一步地，画像服务器30还可以进一步计算用户的群体标签与个体标签之间的关联度。例如，用户的个体标签为“网购”和“游戏”，画像服务器30为用户生成的群体标签为“宅”，此时，画像服务器30还可以进一步计算“宅”这一群体标签分别与“网购”和“游戏”之间的关联度。那么，后续画像优化模块204还可以根据上述关联度对“网购”和“游戏”这两个个体标签的特征值进行校正，从而提高最终生成的用户画像的准确度。
画像查询模块205
画像查询模块205用于响应应用层中任意应用查询用户画像的请求。示例性的，画像查询模块205可提供安卓统一标准的Provider接口，应用可通过调用该Provider接口请求第一画像管理模块201向其提供用户画像。
另外，画像查询模块205向应用提供用户画像时，还可通过数字签名等方式对请求提供用户画像的用户身份进行鉴权，以降低用户隐私泄露的风险。
图8为本发明实施例提供的一种画像服务器的架构示意图。如图7所示，画像服务器30可以包括第二画像管理模块301和第二画像计算模块302。
第二画像管理模块301
与上述终端10中的第一画像管理模块201类似的，第二画像管理模块301是画像服务器30中提供用户画像服务的控制中心，第二画像管理模块301与第二画像计算模块302相连。
具体的，第二画像管理模块301可用于接收终端10发送的用户的个体标签，并指示第二画像计算模块302根据不同终端10发送的不同用户的个体标签，计算每个用户的群体标签。当然，第二画像管理模块301还可将生成的不同用户的群体标签发送给终端10，或保存在画像服务器30的数据库(例如HBase等分布式数据库)中。
第二画像计算模块302
与终端10的第一画像计算模块203类似的，第二画像计算模块302中也可包括一系列用于生成群体标签的算法或模型。
示例性的，如图9A所示，第二画像计算模块302可以将具有共性的多个个体标签抽象为一个群体标签。因此，如图9B所示，第二画像计算模块302可以根据多个用户的个体标签，按照上述算法或模型，通过聚类、组合以及特征转换等方法将某一方面具有共性的多个用户划分为一个群体，该群体的群体标签可以作为群体内用户的群体标签。
进一步地，以用户A为例，第一画像计算模块203为用户A生成用户A的群体标签之后，还可以通过机器学习或大数据挖掘等方法进一步确定用户A的群体标签与个体标签之间的关联度，以便后续终端10可以根据该关联度校正用户A的个体标签的特征值。
示例性的，如图10所示，画像服务器30接收终端1为用户A确定出的3个个体标签(P1-P3)、终端2为用户B确定出的3个个体标签(Q1-Q3)、终端3为用户C确定出的3个个体标签(W1-W3)。那么，第二画像计算模块302通过对这些个体标签进行聚类，可确定用户A属于“90后”这一群体标签的群体，同时，通过对这些个体标签进行特征组合后再进行聚类，可确定用户A属于“美剧”这一群体标签的群体，并且，通过对这些个体标签进行特征转换后再进行聚类，可确定用户A属于“游戏”这一群体标签的群体。
那么，第二画像计算模块302得到用户A的三个群体标签为“90后”(S1)、“美剧”(S2)以及“游戏”(S3)。此时，第二画像计算模块302可继续对用户A的群体标签进行大数据统计或数据挖掘，计算用户A的每一个群体标签与每一个个体标签之间的关联度。例如，群体标签“90后”(S1)与个体标签“吃货”(P1)之间的关联度为90分(以满分100分为例)，说明当用户A为“90后”时，其具有“吃货”这一特征的几率为90％左右，那么，后续终端10可根据该关联度校正终端原本生成的个体标签“吃货”(P1)的特征值。
可以看出，第二画像计算模块302可基于多个用户的个体标签确定出每个用户的群体标签，即用户的群体属性，这样，终端10既可以生成用户的个体标签，又可以获取到用户的群体标签，从而生成更加完整、准确的用户画像。
另外，第二画像计算模块302还可以计算出用户的群体标签与个体标签之间的关联度，使得终端10后续可以对已生成的个体标签的特征值进行校准，进一步提高最终 生成的用户画像的准确度，从而提高终端10提供智慧化服务时的准确度和智能度。
图11为本发明实施例提供的一种用户画像的生成方法的交互示意图。该方法应用于上述终端10和画像服务器30组成的画像***中。如图11所示，该方法包括：
S1001、终端采集目标用户使用终端时产生的行为数据。
具体的，参见对上述终端内数据采集模块202的相关描述，可由数据采集模块202通过***监听、读取特定数据接口、调用***服务、打点采集等方式中的一种或多种，采集目标用户(例如用户A)使用该终端时产生的行为数据，例如，该行为数据具体可以包括应用层级数据、***层级数据以及传感器层级数据。
具体的，对于不同类型的行为数据终端可以设置不同的采集周期。示例性的，对于涉及频繁用户操作的应用或功能，终端可以设置较小的采集周期采集用户的行为数据。例如，终端可每隔5分钟采集终端的位置信息、蓝牙的工作状态等。而对于涉及用户操作不太频繁的应用或功能，终端可以设置较大的采集周期采集用户的行为数据。例如，终端可每隔24小时钟采集终端内安装的应用的名称和数量。
进一步地，数据采集模块202可以将采集到的行为数据存储在终端的数据库(例如SQLite)中，例如，以列表的形式将采集时间以及与采集时间对应行为数据之间的对应关系存储在终端的数据库中。另外，在存储该行为数据时，终端还可以使用加密算法(例如AES256)对采集到的行为数据进行加密处理。
S1002、终端根据上述行为数据为目标用户生成个体标签，该个体标签反映了目标用户的个体行为特征。
在步骤S1002中，在采集到用户的行为数据后，终端内的第一画像管理模块201可将一定时间内采集到的行为数据输入至第一画像计算模块203，由第一画像计算模块203按照预先存储的算法或模型，通过机器学习或统计分析等方法基于采集到的行为数据确定出能够反映用户A行为特征的个体标签。
例如，第一画像管理模块201向第一画像计算模块203发送的行为数据为：最近24小时内采集到的拍照数量。那么，当该拍照数量大于第一预设值(例如15张)时，第一画像计算模块203可以将“爱摄影”确定为用户的用户标签之一，此时对应的特征值为60分(以满分为100举例)；当该拍照数量大于第二预设值(例如25张，第二预设值大于第一预设值)时，第一画像计算模块203可以将“爱摄影”确定为用户的用户标签之一，此时对应的特征值为80分。
当然，第一画像管理模块201还可以使用其他算法或模型生成目标用户的个体标签，例如，排序、加权以及平均、逻辑回归算法、Adaboost算法、朴素贝叶斯算法以及神经网络算法等，本申请实施例对此不做任何限制。另外，第一画像计算模块203为目标用户确定出的个人标签可以包括一个或多个，本申请实施例对此不做任何限制。
S1003、终端将上述个体标签中敏感度小于阈值的个体标签发送至画像服务器。
在步骤S1003中，终端可对步骤S1002生成的涉及目标用户隐私的个体标签(例如用户A的住址、电话等)进行脱敏处理，使得生成的个体标签中尽可能少的存在涉及用户隐私的个体标签。
经脱敏处理后，终端可分别确定每个个体标签与用户隐私之间的相关程度，例如， 通过计算个体标签与用户隐私之间的置信度或相关系数等方法，得到该个体标签的敏感度。
当个体标签与用户隐私之间的相关程度越大，则该个体标签的敏感度越大，相应的，当个体标签与用户隐私之间的相关程度越小，则该个体标签的敏感度越小。那么，当个体标签的敏感度小于阈值时，说明该个体标签所反映的用户隐私较少，因此，终端可将敏感度小于上述阈值的一个或多个目标用户的个体标签发送至画像服务器，降低终端与画像服务器交互时泄露用户隐私的风险。
S1004、画像服务器获取N个用户中每个用户的个体标签，N＞1。
其中，上述N个用户中包括步骤S1001-S1003中所述的目标用户。
由于每个用户使用终端时，其终端均可通过执行上述步骤S1001-S1003向画像服务器发送为各个用户生成的个体标签，因此，画像服务器每接收到一个终端向其发送的个体标签时，均可将其存储在画像服务器的数据库中，从而得到N个用户中每个用户的个体标签。
S1005、画像服务器根据N个用户中每个用户的个体标签，生成目标用户的群体标签，该群体标签反映了目标用户所属的群体的行为特征。
具体的，可参见图9-图10的相关描述，画像服务器的第二画像管理模块301可向第二画像计算模块302输入上述N个用户的个体标签，由第二画像计算模块302按照预设的算法或模型，通过聚类、特征组合以及特征转换等方法确定出N个用户中每个用户的群体标签。
其中，聚类是指将个体标签相似的用户聚合为一类群体。例如，画像服务器预先设置了“90后”这一群体标签与群体1之间的对应关系，该群体1是指包含“爱摄影”和“网购”等个体标签的用户。那么，当第二画像管理模块301检测到用户A与用户B均具有“爱摄影”和“网购”的个体标签时，可将用户A与用户B作为群体1中的两位成员。由于群体1的群体标签为“90后”，因此，属于群体1的用户A和用户B的群体标签也包括“90后”。
特征组合是指将数量较多的个体标签按照一定规则组合后转换为数量较少的特征标签。后续，画像服务器可使用这些特征标签进一步通过上述聚类算法将相似的用户聚合为一类群体。例如，用户A的个体标签一共有50个，那么，第二画像管理模块301可以按照衣、食、住、行这4种类型将这50个个体标签组合为4个特征标签。后续，第二画像管理模块301可以将用户A、用户B以及用户C等用户分别在衣、食、住、行这4种类型的4个特征标签进行聚类，从而得到每个用户的群体标签。
特征转换是指将用户的多个个体标签分别转换为对应的转换标签，例如，用户A的个体标签为“QQ长时间在线”、“转账频率高”以及“机票”，那么，可将“QQ长时间在线”转换为“网聊”，将“转账频率高”转换为“高收入”，将“机票”转换为“出行”。进而，第二画像管理模块301可使用“网聊”、“高收入”以及“出行”这三个转换标签与其他用户特征转换后的转换标签进行聚类，从而得到每个用户的群体标签。
当然，画像服务器具体可以使用逻辑回归算法、Adaboost算法、规约映射算法、回归分析算法、Web数据挖掘算法、随机森林(Random Forests)算法以及K-最近邻 法(K-nearest neighbors)等算法将多个用户归属在不同特征的群体中，并从而赋予不同用户相应的群体标签，本申请实施例对此不做任何限制。
也就是说，画像服务器可基于多个用户的个体标签，综合分析出目标用户所属的群体，从而得到目标用户的群体标签，该群体标签可反映出用户所属群体的行为特征，这样，终端后续结合目标用户的个体标签和群体标签，可以得到更加完整准确的用户画像。
需要说明的是，第二画像管理模块301为目标用户确定出的群体标签可以包括一个或多个，本申请实施例对此不做任何限制。
另外，步骤1005中画像服务器根据N个用户中每个用户的个体标签，生成目标用户的群体标签可以是一个不断循环的过程。例如，当画像服务器接收到新用户发来的个体标签后，可以将该新用户的个人标签作为新的输入，结合图10中用户A-用户C的个体标签重新确定每个用户所属的群体以及每个用户的群体标签。也就是说，画像服务器为用户A确定的群体标签可以是一个不断更新的过程，并且，画像服务器可以将每次更新后的用户A的群体标签发送给用户A的终端。
S1006(可选的)、画像服务器确定目标用户的群体标签与个体标签之间的关联度。
可选的，在步骤S1006中，画像服务器得到目标用户的群体标签后，可进一步对该群体标签进行大数据下的关联规则(Association Rule)挖掘，以确定目标用户的群体标签与目标用户的个体标签之间的关联度。
例如，群体标签“90后”与个体标签“熬夜”之间的关联度为90(以满分100举例)，即说明当目标用户具有“90后”这一群体标签时，有90％左右的概率会出现“熬夜”这一个体标签。那么，后续终端可根据画像服务器确定出的关联度优化步骤S1002中生成的个体标签及个体标签的特征值，从而提高最终生成的用户画像的准确率。
S1007、画像服务器将目标用户的群体标签发送给终端。
需要说明的是，当画像服务器执行上述步骤S1006得到目标用户的群体标签与个体标签之间的关联度后，还可以在步骤S1007中同时将上述关联度发送给终端。
S1008、终端根据目标用户的群体标签以及上述关联度，对步骤S1002生成的个体标签进行校正，得到目标用户的用户画像。
在步骤S1007-S1008中，画像服务器将步骤S1005和S1006中得到的目标用户的群体标签以及上述关联度发送给终端，使得终端可以根据该群体标签和关联度对步骤S1002中基于行为数据生成的个体标签进行校正，为目标用户生成最终的用户画像。
示例性的，终端在执行步骤S1002时为用户生成的个体标签包括两部分，一部分是与用户隐私关联度较高的个体标签，这些个体标签没有发送给画像服务器；另一部分是与用户隐私关联度较低的个体标签，这些个体标签发送给画像服务器后，由画像服务器为目标用户生成了群体标签。
那么，如图12所示，对于画像服务器发送的目标用户的群体标签，由于终端在步骤S1002中生成目标用户的个体标签时没有考虑目标用户的群体特征，因此，终端可将目标用户的群体标签以及原本生成的个体标签的全集作为目标用户的用户画像。
又或者，终端还可以将目标用户的群体标签作为新的行为数据输入画像优化模块 204，由画像优化模块204结合目标用户的行为数据和群体标签重新计算目标用户的个体标签。这样生成的个体标签综合考虑了目标用户的个体行为特征和目标用户所属的群体行为特征，使得优化后的个体标签更加完整和准确。
进一步地，仍如图12所示，对于画像服务器发送的群体标签与个体标签之间的关联度，画像优化模块204可以根据该关联度对步骤S1002中生成的个体标签进行校正。
例如，画像服务器发送的群体标签“90后”与个体标签“网购”之间的关联度为75，说明当目标用户具有群体标签“90后”时，同时具有个体标签“网购”的几率大约为75％，而步骤S1002中生成的个体标签并不包括“网购”这一个体标签，则终端可在目标用户的个体标签中添加“网购”这一标签，并设置“网购”这一个体标签的特征值，该特征值可以为小于75分的任意值。
又例如，画像服务器发送的群体标签“90后”与个体标签“熬夜”之间的关联度为95，说明当目标用户具有群体标签“90后”时，同时具有个体标签“熬夜”的几率大约为95％，而步骤S1002中终端为个体标签“熬夜”的特征值为65分，说明终端判断出的“熬夜”这一特征的特征值可能有偏差。因此，终端可在原始特征值65分的基础上根据上述关联度(95)校正对“熬夜”这一个体标签的特征值。
例如，校正后的特征值＝原始特征值+关联度*校正因子。
其中，校正因子可以为-1到1之间的任意数值，校正因子的大小可以反映出关联度对个体标签特征值的影响程度。以校正因子为0.1举例，对于“熬夜”这一个体标签，其校正后的特征值＝原始特征值(65)+关联度(95)*校正因子(0.1)＝74.5。
这样，终端可以将校正后得到的个体标签以及个体标签的特征值作为目标用户的用户画像，此时得到的用户画像不仅综合考虑了用户个体行为特征和群体行为特征的影响，还对用户画像中个体标签的特征值进行了校正，使得终端最终生成的目标用户的用户画像的完整性和准确率得到提高。
S1009、当终端接收到第一应用获取用户画像的请求时，终端向第一应用提供上述用户画像。
在步骤S1009中，终端为目标用户生成准确度更高、完整性更强的用户画像后，如果检测到终端上运行的应用(例如第一应用)需要向用户提供智慧化业务时，第一应用可通过调用画像查询模块205中的Provider等特定接口，请求第一画像管理模块201向第一应用提供用户画像，此时，第一画像管理模块201可将步骤S1008中生成的用户画像作为请求结果反馈给第一应用。
由于上述用户画像是终端与画像服务器协同交互后生成的完整性和准确率较高的用户画像，因此，第一应用使用该用户画像可为用户提供更加智能、便捷的智慧化业务。
其中，上述步骤S1001-S1003以及S1008-S1009中涉及终端的执行步骤，可以由图1所示的终端的处理器执行其存储器中存储的程序指令来实现。类似的，上述步骤S1004-S1007中涉及画像服务器的执行步骤，可以由画像服务器的处理器执行其存储器中存储的程序指令来实现。
可以理解的是，上述终端等为了实现上述功能，其包含了执行各个功能相应的硬 件结构和/或软件模块。本领域技术人员应该很容易意识到，结合本文中所公开的实施例描述的各示例的单元及算法步骤，本申请实施例能够以硬件或硬件和计算机软件的结合形式来实现。某个功能究竟以硬件还是计算机软件驱动硬件的方式来执行，取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能，但是这种实现不应认为超出本申请实施例的范围。
本申请实施例可以根据上述方法示例对上述终端等进行功能模块的划分，例如，可以对应各个功能划分各个功能模块，也可以将两个或两个以上的功能集成在一个处理模块中。上述集成的模块既可以采用硬件的形式实现，也可以采用软件功能模块的形式实现。需要说明的是，本申请实施例中对模块的划分是示意性的，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式。
在采用对应各个功能划分各个功能模块的情况下，图3示出了上述实施例中所涉及的终端的一种可能的结构示意图，包括：第一画像管理模块201、数据采集模块202、第一画像计算模块203、第一画像查询模块205以及画像优化模块204。其中，这些功能模块的相关动作均可以援引到图3的相关描述中，在此不再赘述。
在采用对应各个功能划分各个功能模块的情况下，图8示出了上述实施例中所涉及的画像服务器的一种可能的结构示意图，包括：第二画像管理模块301和第二画像计算模块302。其中，这些功能模块的相关动作均可以援引到图8的相关描述中，在此不再赘述。
在采用集成的单元的情况下，如图13所示，示出了上述实施例中所涉及的终端的一种可能的结构示意图，包括处理模块2101、通信模块2102、输入/输出模块2103以及存储模块2104。
其中，处理模块2101用于对终端的动作进行控制管理。通信模块2102用于支持终端与其他网络实体的通信。输入/输出模块2103用于接收由用户输入的信息或输出提供给用户的信息以及终端的各种菜单。存储模块2104用于保存终端的程序代码和数据。
在采用集成的单元的情况下，如图14所示，示出了上述实施例中所涉及的画像服务器的一种可能的结构示意图，包括处理模块2201、通信模块2202以及存储模块2203。
其中，处理模块2201用于对画像服务器的动作进行控制管理。通信模块2202用于支持画像服务器与其他服务器或终端的通信。存储模块2203用于保存画像服务器的程序代码和数据。
具体的，上述处理模块2101/2201可以是处理器或控制器，例如可以是中央处理器(Central Processing Unit，CPU)，GPU，通用处理器，数字信号处理器(Digital Signal Processor，DSP)，专用集成电路(Application-Specific Integrated Circuit，ASIC)，现场可编程门阵列(Field Programmable Gate Array，FPGA)或者其他可编程逻辑器件、晶体管逻辑器件、硬件部件或者其任意组合。其可以实现或执行结合本申请公开内容所描述的各种示例性的逻辑方框，模块和电路。所述处理器也可以是实现计算功能的组合，例如包含一个或多个微处理器组合，DSP和微处理器的组合等等。
上述通信模块2102/2202可以是收发器、收发电路、或通信接口等。例如，通信模块1303具体可以是蓝牙装置、Wi-Fi装置、外设接口等等。
上述输入/输出模块2103可以是触摸屏、显示器、麦克风等接收用户输入的信息或输出向用户提供的信息的设备。以显示器为例，具体可以采用液晶显示器、有机发光二极管等形式来配置显示器。另外，显示器上还可以集成触控板，用于采集在其上或附近的触摸事件，并将采集到的触摸信息发送给其他器件(例如处理器等)。
上述存储模块2104/2203可以是存储器，该存储器可以包括高速随机存取存储器(RAM)，还可以包括非易失存储器，例如磁盘存储器件、闪存器件或其他易失性固态存储器件等。
在上述实施例中，可以全部或部分的通过软件，硬件，固件或者其任意组合来实现。当使用软件程序实现时，可以全部或部分地以计算机程序产品的形式出现。所述计算机程序产品包括一个或多个计算机指令。在计算机上加载和执行所述计算机程序指令时，全部或部分地产生按照本申请实施例所述的流程或功能。所述计算机可以是通用计算机、专用计算机、计算机网络、或者其他可编程装置。所述计算机指令可以存储在计算机可读存储介质中，或者从一个计算机可读存储介质向另一个计算机可读存储介质传输，例如，所述计算机指令可以从一个网站站点、计算机、服务器或数据中心通过有线(例如同轴电缆、光纤、数字用户线(DSL))或无线(例如红外、无线、微波等)方式向另一个网站站点、计算机、服务器或数据中心传输。所述计算机可读存储介质可以是计算机能够存取的任何可用介质或者是包含一个或多个可用介质集成的服务器、数据中心等数据存储设备。该可用介质可以是磁性介质，(例如，软盘，硬盘、磁带)、光介质(例如，DVD)或者半导体介质(例如固态硬盘Solid State Disk(SSD))等。
以上所述，仅为本申请的具体实施方式，但本申请的保护范围并不局限于此，任何在本申请揭露的技术范围内的变化或替换，都应涵盖在本申请的保护范围之内。因此，本申请的保护范围应以所述权利要求的保护范围为准。

Claims (30)

  1. 一种用户画像的生成方法，其特征在于，包括：
    终端将为用户生成的至少一个个体标签发送至画像服务器，所述个体标签反映了所述用户的个人行为特征；
    所述终端接收所述画像服务器为所述用户生成的至少一个群体标签，所述群体标签是所述画像服务器至少基于所述至少一个个体标签生成的，所述群体标签反映了所述用户所属群体的行为特征；
    所述终端使用所述群体标签更新所述用户的用户画像；
    所述终端向第一应用提供更新后的用户画像中的至少一部分。
  2. 根据权利要求1所述的方法，其特征在于，在终端将为用户生成的至少一个个体标签发送至画像服务器之前，还包括：
    所述终端采集所述用户使用所述终端时产生的行为数据；
    所述终端根据所述行为数据为所述用户生成至少一个个体标签，每个个体标签中包括该个体标签的类型和该个体标签的特征值。
  3. 根据权利要求1或2所述的方法，其特征在于，终端将为用户生成的至少一个个体标签发送至画像服务器，包括：
    所述终端将所述至少一个个体标签中敏感度小于阈值的个体标签发送至所述画像服务器，所述敏感度用于指示所述个体标签与用户隐私之间的相关程度。
  4. 根据权利要求1-3中任一项所述的方法，其特征在于，所述终端使用所述群体标签更新所述用户的用户画像，包括：
    所述终端将所述群体标签添加至所述用户的用户画像中，得到更新后的用户画像，所述更新后的用户画像中包括所述群体标签和所述终端生成的至少一个个体标签。
  5. 根据权利要求2或3所述的方法，其特征在于，所述终端使用所述群体标签更新所述用户的用户画像，包括：
    所述终端根据所述行为数据和所述群体标签更新所述终端生成的至少一个个体标签，得到更新后的用户画像，所述更新后的用户画像中包括所述更新后的个体标签。
  6. 根据权利要求5所述的方法，其特征在于，所述更新后的用户画像中还包括所述群体标签。
  7. 根据权利要求2-6中任一项所述的方法，其特征在于，在所述终端接收所述画像服务器为所述用户生成的至少一个群体标签之后，还包括：
    所述终端接收所述画像服务器为所述用户生成的第一群体标签与第一个体标签之间的关联度，所述第一群体标签为所述至少一个群体标签中的一个，所述第一个体标签为所述至少一个个体标签中的一个；
    所述终端根据所述关联度校正所述第一个体标签的特征值。
  8. 根据权利要求7所述的方法，其特征在于，所述终端根据所述关联度校正所述第一个体标签的特征值，包括：
    所述终端将所述第一个体标签的特征值与校正值之和作为所述第一个体标签校正后的特征值，所述校正值为所述关联度与预设的校正因子的乘积，所述校正因子用于反映所述关联度对所述第一个体标签的影响程度。
  9. 一种用户画像的生成方法，其特征在于，包括：
    画像服务器获取至少一个用户的个体标签；
    所述画像服务器根据所述至少一个用户中每个用户的个体标签，生成目标用户的群体标签，所述群体标签反映了所述目标用户所属群体的行为特征，所述目标用户为所述至少一个用户中的一个；
    所述画像服务器将所述目标用户的群体标签发送给终端。
  10. 根据权利要求9所述的方法，其特征在于，所述画像服务器根据所述至少一个用户中每个用户的个体标签，生成目标用户的群体标签，包括：
    所述画像服务器根据所述至少一个用户中每个用户的个体标签，将所述至少一个用户划分为至少一个群体；
    所述画像服务器将所述目标用户所属群体的标签作为所述目标用户的群体标签。
  11. 根据权利要求10所述的方法，其特征在于，所述画像服务器根据所述至少一个用户中每个用户的个体标签，将所述至少一个用户划分为至少一个群体，包括：
    所述画像服务器基于所述至少一个用户中每个用户的个体标签，通过聚类、特征组合后聚类以及特征转换后聚类中的一种或多种方式，将所述至少一个用户划分为至少一个群体。
  12. 根据权利要求9-11中任一项所述的方法，其特征在于，在画像服务器获取至少一个用户的个体标签之后，还包括：
    所述画像服务器确定所述目标用户的群体标签分别与所述目标用户的每个个体标签之间的关联度；
    所述画像服务器将所述关联度发送给所述终端。
  13. 一种终端，其特征在于，包括画像管理模块，以及与所述画像管理模块均相连的数据采集模块、画像计算模块、画像优化模块以及画像查询模块，其中，
    所述画像管理模块，用于：将为用户生成的至少一个个体标签发送至画像服务器，所述个体标签反映了所述用户的个人行为特征；接收所述画像服务器为所述用户生成的至少一个群体标签，所述群体标签是所述画像服务器至少基于所述至少一个个体标签生成的，所述群体标签反映了所述用户所属群体的行为特征；
    所述画像优化模块，用于：使用所述群体标签更新所述用户的用户画像；
    所述画像查询模块，用于：向第一应用提供更新后的所述用户画像中的至少一部分。
  14. 根据权利要求13所述的终端，其特征在于，
    所述数据采集模块，用于：采集所述用户使用所述终端时产生的行为数据；
    所述画像计算模块，用于：根据所述行为数据为所述用户生成至少一个个体标签，每个个体标签中包括该个体标签的类型和该个体标签的特征值。
  15. 根据权利要求13或14所述的终端，其特征在于，
    所述画像管理模块，具体用于：将所述至少一个个体标签中敏感度小于阈值的个体标签发送至所述画像服务器，所述敏感度用于指示所述个体标签与用户隐私之间的相关程度。
  16. 根据权利要求13-15中任一项所述的终端，其特征在于，
    所述画像优化模块，具体用于：将所述群体标签添加至所述用户的用户画像中，得到更新后的用户画像，所述更新后的用户画像中包括所述群体标签和所述终端生成的至少一个个体标签。
  17. 根据权利要求14或15所述的终端，其特征在于，
    所述画像优化模块，具体用于：根据所述行为数据和所述群体标签更新所述终端生成的至少一个个体标签，得到更新后的用户画像，所述更新后的用户画像中包括所述更新后的个体标签。
  18. 根据权利要求17所述的终端，其特征在于，所述更新后的用户画像中还包括所述群体标签。
  19. 根据权利要求14-18中任一项所述的终端，其特征在于，
    所述画像管理模块，还用于：接收所述画像服务器为所述用户生成的第一群体标签与第一个体标签之间的关联度，所述第一群体标签为所述至少一个群体标签中的一个，所述第一个体标签为所述至少一个个体标签中的一个；
    所述画像优化模块，还用于：根据所述关联度校正所述第一个体标签的特征值。
  20. 根据权利要求19所述的终端，其特征在于，
    所述画像优化模块，具体用于：将所述第一个体标签的特征值与校正值之和作为所述第一个体标签校正后的特征值，所述校正值为所述关联度与预设的校正因子的乘积，所述校正因子用于反映所述关联度对所述第一个体标签的影响程度。
  21. 一种服务器，其特征在于，包括画像管理模块，以及与所述画像管理模块相连的画像计算模块，其中，
    所述画像管理模块，用于：获取至少一个用户的个体标签；
    所述画像计算模块，用于：根据所述至少一个用户中每个用户的个体标签，生成目标用户的群体标签，所述群体标签反映了所述目标用户所属群体的行为特征，所述目标用户为所述至少一个用户中的一个；
    所述画像管理模块，还用于：将所述目标用户的群体标签发送给终端。
  22. 根据权利要求21所述的服务器，其特征在于，
    所述画像计算模块，具体用于：根据所述至少一个用户中每个用户的个体标签，将所述至少一个用户划分为至少一个群体；将所述目标用户所属群体的标签作为所述目标用户的群体标签。
  23. 根据权利要求22所述的服务器，其特征在于，
    所述画像计算模块，具体用于：基于所述至少一个用户中每个用户的个体标签，通过聚类、特征组合后聚类以及特征转换后聚类中的一种或多种方式，将所述至少一个用户划分为至少一个群体。
  24. 根据权利要求21-23中任一项所述的服务器，其特征在于，
    所述画像计算模块，还用于：确定所述目标用户的群体标签分别与所述目标用户的每个个体标签之间的关联度；
    所述画像管理模块，还用于：将所述关联度发送给所述终端。
  25. 一种终端，其特征在于，包括：处理器、存储器、总线和通信接口；
    所述存储器用于存储计算机执行指令，所述处理器与所述存储器通过所述总线连 接，当所述终端运行时，所述处理器执行所述存储器存储的所述计算机执行指令，以使所述终端执行如权利要求1-8中任一项所述的用户画像的生成方法。
  26. 一种画像服务器，其特征在于，包括：处理器、存储器、总线和通信接口；
    所述存储器用于存储计算机执行指令，所述处理器与所述存储器通过所述总线连接，当所述画像服务器运行时，所述处理器执行所述存储器存储的所述计算机执行指令，以使所述画像服务器执行如权利要求9-12中任一项所述的用户画像的生成方法。
  27. 一种计算机可读存储介质，所述计算机可读存储介质中存储有指令，其特征在于，当所述指令在终端上运行时，使得所述终端执行如权利要求1-8中任一项所述的用户画像的生成方法。
  28. 一种计算机可读存储介质，所述计算机可读存储介质中存储有指令，其特征在于，当所述指令在画像服务器上运行时，使得所述画像服务器执行如权利要求9-12中任一项所述的用户画像的生成方法。
  29. 一种包含指令的计算机程序产品，其特征在于，当所述计算机程序产品在终端上运行时，使得所述终端执行如权利要求1-8中任一项所述的用户画像的生成方法。
  30. 一种包含指令的计算机程序产品，其特征在于，当所述计算机程序产品在画像服务器上运行时，使得所述画像服务器执行如权利要求9-12中任一项所述的用户画像的生成方法。
PCT/CN2018/073673 2018-01-22 2018-01-22 一种用户画像的生成方法及装置 WO2019140703A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201880019023.3A CN110431585B (zh) 2018-01-22 2018-01-22 一种用户画像的生成方法及装置
PCT/CN2018/073673 WO2019140703A1 (zh) 2018-01-22 2018-01-22 一种用户画像的生成方法及装置

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2018/073673 WO2019140703A1 (zh) 2018-01-22 2018-01-22 一种用户画像的生成方法及装置

Publications (1)

Publication Number Publication Date
WO2019140703A1 true WO2019140703A1 (zh) 2019-07-25

Family

ID=67301981

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/073673 WO2019140703A1 (zh) 2018-01-22 2018-01-22 一种用户画像的生成方法及装置

Country Status (2)

Country Link
CN (1) CN110431585B (zh)
WO (1) WO2019140703A1 (zh)

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110445937A (zh) * 2019-09-16 2019-11-12 Oppo（重庆）智能科技有限公司 事件提醒方法及相关产品
CN110991392A (zh) * 2019-12-17 2020-04-10 Oppo广东移动通信有限公司 人群识别方法、装置、终端及存储介质
WO2021068608A1 (zh) * 2019-10-11 2021-04-15 深圳壹账通智能科技有限公司 提取用户画像的方法、装置、计算机设备和存储介质
CN113055212A (zh) * 2019-12-27 2021-06-29 Oppo广东移动通信有限公司 策略的推送方法、策略的执行方法、装置、设备及介质
CN113051465A (zh) * 2019-12-27 2021-06-29 Oppo广东移动通信有限公司 优化策略的推送方法、装置、服务器及存储介质
CN113051464A (zh) * 2019-12-27 2021-06-29 Oppo广东移动通信有限公司 策略的推送方法、策略的执行方法、装置、设备及介质
WO2021142719A1 (zh) * 2020-01-16 2021-07-22 深圳市欢太科技有限公司 画像生成方法、装置、服务器及存储介质
CN114218241A (zh) * 2021-12-17 2022-03-22 福建凯米网络科技有限公司 一种用户画像更新方法、装置及存储介质
CN114693245A (zh) * 2022-03-02 2022-07-01 深圳市小满科技有限公司 用户画像生成方法及装置、电子设备、可读存储介质
CN115004210A (zh) * 2020-04-01 2022-09-02 深圳市欢太科技有限公司 用户画像列表构建方法、装置、服务器以及存储介质
CN115062725A (zh) * 2022-07-12 2022-09-16 北京威控科技股份有限公司 酒店收益异常分析方法及***
WO2024046081A1 (zh) * 2022-08-30 2024-03-07 华为技术有限公司 信息推荐方法、电子设备及服务器

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113050961A (zh) * 2019-12-27 2021-06-29 Oppo广东移动通信有限公司 优化策略的推送方法、装置、服务器及存储介质
CN113806656B (zh) * 2020-06-17 2024-04-26 华为技术有限公司 确定用户的特征的方法、设备及计算机可读介质
CN111967970B (zh) * 2020-08-18 2024-02-27 中国银行股份有限公司 基于spark平台的银行产品推荐方法及装置
CN112465565B (zh) * 2020-12-11 2023-09-26 加和（北京）信息科技有限公司 一种基于机器学习的用户画像预测的方法及装置
CN112561598A (zh) * 2020-12-23 2021-03-26 中国农业银行股份有限公司重庆市分行 基于客户画像的客户流失预测及挽回方法和***
CN114627991B (zh) * 2022-03-03 2022-10-14 元盛视光(湖北)生物科技有限公司 一种基于智慧医疗大数据的用户隐私保护方法及***

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103577549A (zh) * 2013-10-16 2014-02-12 复旦大学 一种基于微博标签的人群画像***和方法
US20160063536A1 (en) * 2014-08-27 2016-03-03 InMobi Pte Ltd. Method and system for constructing user profiles
CN105893407A (zh) * 2015-11-12 2016-08-24 乐视云计算有限公司 个体用户画像方法和***
CN106354755A (zh) * 2016-08-17 2017-01-25 洑云龙 一种用户画像优化处理方法
CN106373026A (zh) * 2016-08-24 2017-02-01 国网冀北电力有限公司电力科学研究院 一种电力行业的用户画像构建方法

Family Cites Families (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8055664B2 (en) * 2007-05-01 2011-11-08 Google Inc. Inferring user interests
US8543518B2 (en) * 2010-07-26 2013-09-24 Yahoo! Inc. Deducing shadow user profiles for ad campaigns
EP2862102A4 (en) * 2012-06-14 2016-01-27 Nokia Technologies Oy METHOD AND APPARATUS FOR ASSOCIATING LABELS OF INTEREST WITH MULTIMEDIA ELEMENTS BASED ON SOCIAL DIFFUSIONS BETWEEN USERS
CN105677647B (zh) * 2014-11-17 2019-04-19 ***通信集团广东有限公司 一种个体推荐方法及***
CN106503015A (zh) * 2015-09-07 2017-03-15 国家计算机网络与信息安全管理中心 一种构建用户画像的方法
CN105893406A (zh) * 2015-11-12 2016-08-24 乐视云计算有限公司 群体用户画像方法及***
US20170142119A1 (en) * 2015-11-12 2017-05-18 Le Holdings (Beijing) Co., Ltd. Method for creating group user profile, electronic device, and non-transitory computer-readable storage medium
CN106339433B (zh) * 2016-08-18 2021-08-24 冯连元 一种基于数据中有关群体数据与个体数据交互比较平台的方法及装置
CN106933946A (zh) * 2017-01-20 2017-07-07 深圳市三体科技有限公司 一种基于移动终端的大数据管理方法及***
CN106909686A (zh) * 2017-03-06 2017-06-30 吉林省盛创科技有限公司 一种人机交互构建用户画像聚类计算方法
CN107464142B (zh) * 2017-08-09 2018-11-09 南京甄视智能科技有限公司 基于营销反馈结果反哺用户画像***的方法、用户画像***

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103577549A (zh) * 2013-10-16 2014-02-12 复旦大学 一种基于微博标签的人群画像***和方法
US20160063536A1 (en) * 2014-08-27 2016-03-03 InMobi Pte Ltd. Method and system for constructing user profiles
CN105893407A (zh) * 2015-11-12 2016-08-24 乐视云计算有限公司 个体用户画像方法和***
CN106354755A (zh) * 2016-08-17 2017-01-25 洑云龙 一种用户画像优化处理方法
CN106373026A (zh) * 2016-08-24 2017-02-01 国网冀北电力有限公司电力科学研究院 一种电力行业的用户画像构建方法

Cited By (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110445937A (zh) * 2019-09-16 2019-11-12 Oppo（重庆）智能科技有限公司 事件提醒方法及相关产品
WO2021068608A1 (zh) * 2019-10-11 2021-04-15 深圳壹账通智能科技有限公司 提取用户画像的方法、装置、计算机设备和存储介质
CN110991392A (zh) * 2019-12-17 2020-04-10 Oppo广东移动通信有限公司 人群识别方法、装置、终端及存储介质
CN113055212B (zh) * 2019-12-27 2022-11-15 Oppo广东移动通信有限公司 策略的推送方法、策略的执行方法、装置、设备及介质
CN113055212A (zh) * 2019-12-27 2021-06-29 Oppo广东移动通信有限公司 策略的推送方法、策略的执行方法、装置、设备及介质
CN113051465A (zh) * 2019-12-27 2021-06-29 Oppo广东移动通信有限公司 优化策略的推送方法、装置、服务器及存储介质
CN113051464A (zh) * 2019-12-27 2021-06-29 Oppo广东移动通信有限公司 策略的推送方法、策略的执行方法、装置、设备及介质
WO2021142719A1 (zh) * 2020-01-16 2021-07-22 深圳市欢太科技有限公司 画像生成方法、装置、服务器及存储介质
CN115004210A (zh) * 2020-04-01 2022-09-02 深圳市欢太科技有限公司 用户画像列表构建方法、装置、服务器以及存储介质
CN114218241A (zh) * 2021-12-17 2022-03-22 福建凯米网络科技有限公司 一种用户画像更新方法、装置及存储介质
CN114693245A (zh) * 2022-03-02 2022-07-01 深圳市小满科技有限公司 用户画像生成方法及装置、电子设备、可读存储介质
CN115062725A (zh) * 2022-07-12 2022-09-16 北京威控科技股份有限公司 酒店收益异常分析方法及***
CN115062725B (zh) * 2022-07-12 2023-08-08 北京威控科技股份有限公司 酒店收益异常分析方法及***
WO2024046081A1 (zh) * 2022-08-30 2024-03-07 华为技术有限公司 信息推荐方法、电子设备及服务器

Also Published As

Publication number Publication date
CN110431585A (zh) 2019-11-08
CN110431585B (zh) 2024-03-15

Similar Documents

Publication Publication Date Title
WO2019140703A1 (zh) 一种用户画像的生成方法及装置
CN110782289B (zh) 一种基于用户画像的业务推荐方法和***
US20200107152A1 (en) Inferring user availability for a communication
US20200204643A1 (en) User profile generation method and terminal
WO2019140702A1 (zh) 一种用户画像的生成方法及装置
US20190057298A1 (en) Mapping actions and objects to tasks
US10664484B2 (en) Computerized system and method for optimizing the display of electronic content card information when providing users digital content
US10129336B2 (en) Content management method and cloud server therefor
US20170364947A1 (en) System and method for event triggered search results
US20120209839A1 (en) Providing applications with personalized and contextually relevant content
US20130110992A1 (en) Electronic device management using interdomain profile-based inferences
KR20140026305A (ko) 친구 추천 방법 및 이를 위한 서버 및 단말
US20210337010A1 (en) Computerized system and method for automatically providing networked devices non-native functionality
US20230206034A1 (en) Prediction of Next Place Visits on Online Social Networks
KR20140113436A (ko) 관계 모델 매커니즘을 구현한 컴퓨팅 시스템 및 그 작동 방법
JP6543774B2 (ja) コンピューティングデバイスのプロパティに基づく情報のランク付け
KR20140027011A (ko) 친구 추천 방법 및 이를 위한 서버 및 단말
US20170249325A1 (en) Proactive favorite leisure interest identification for personalized experiences
US20190180386A1 (en) Cross-Validating Places on Online Social Networks
US10506383B2 (en) Location prediction using wireless signals on online social networks
US20220004703A1 (en) Annotating a collection of media content items
WO2017130727A1 (ja) 情報処理装置および方法、並びにプログラム
US20220417226A1 (en) Automatic privacy-aware machine learning method and apparatus
US20190090197A1 (en) Saving battery life with inferred location
WO2022057764A1 (zh) 广告显示方法及电子设备

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18900793

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18900793

Country of ref document: EP

Kind code of ref document: A1