CN112685516A - Multi-channel recall recommendation method and device, electronic equipment and medium - Google Patents
Multi-channel recall recommendation method and device, electronic equipment and medium Download PDFInfo
- Publication number
- CN112685516A CN112685516A CN202110038094.1A CN202110038094A CN112685516A CN 112685516 A CN112685516 A CN 112685516A CN 202110038094 A CN202110038094 A CN 202110038094A CN 112685516 A CN112685516 A CN 112685516A
- Authority
- CN
- China
- Prior art keywords
- data
- generating
- target
- similarity
- labels
- Prior art date
- Legal status (The legal status 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 status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000007781 pre-processing Methods 0.000 claims abstract description 17
- 238000012163 sequencing technique Methods 0.000 claims abstract description 10
- 238000004364 calculation method Methods 0.000 claims abstract description 7
- 230000006399 behavior Effects 0.000 claims description 38
- 239000013598 vector Substances 0.000 claims description 19
- 238000004422 calculation algorithm Methods 0.000 claims description 16
- 238000012545 processing Methods 0.000 claims description 13
- 238000013527 convolutional neural network Methods 0.000 claims description 4
- 238000001914 filtration Methods 0.000 claims description 4
- 230000004931 aggregating effect Effects 0.000 claims description 3
- 239000000126 substance Substances 0.000 claims description 3
- 238000004590 computer program Methods 0.000 description 13
- 230000008569 process Effects 0.000 description 8
- 238000010586 diagram Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 6
- 230000003287 optical effect Effects 0.000 description 4
- 238000004891 communication Methods 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 2
- 230000000644 propagated effect Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the disclosure provides a multi-channel recall recommendation method, a multi-channel recall recommendation device, electronic equipment and a medium, which belong to the technical field of computing and specifically comprise the following steps: preprocessing the initial data set to obtain a data set to be detected; generating a first label corresponding to all the objects according to the content data, and generating a second label corresponding to the target user according to the historical behavior data; respectively calculating the similarity between the corresponding data of all the first labels; recalling the candidate set corresponding to the target user according to the similarity and the second label; and sequencing the candidate set to obtain a target set. According to the scheme, different types of initial data are processed to obtain classified data, corresponding labels are generated, the similarity between the data corresponding to the first labels is calculated, the candidate sets are recalled in combination with the second labels, the candidate sets are ranked to obtain the target set, and the calculation efficiency and recommendation precision of multi-path recall recommendation are improved.
Description
Technical Field
The present disclosure relates to the field of computing technologies, and in particular, to a multi-way recall recommendation method, apparatus, electronic device, and medium.
Background
At present, the internet can provide massive information for users, and an intelligent recommendation system can help the users to find interesting information quickly. The recommendation algorithm is characterized in that the association relationship between people and articles is established, a core is formed by surrounding data, algorithm and a system, massive data information is applied to a corresponding recall strategy and a corresponding sorting strategy by utilizing the algorithm of the recommendation system, and the personalized recommendation process provided for a user is realized. However, most of the existing recall recommendation methods are processing original data of numerical types, little and few data of texts, audios and pictures are involved, most of algorithms are single-path recall models, and the recommendation efficiency and accuracy are poor.
Therefore, an efficient and accurate multi-recall recommendation algorithm is needed.
Disclosure of Invention
In view of the above, embodiments of the present disclosure provide a method, an apparatus, an electronic device, and a medium for multi-recall recommendation, which at least partially solve the problem of poor efficiency and accuracy of recommendation in the prior art.
In a first aspect, an embodiment of the present disclosure provides a multi-recall recommendation method, including:
preprocessing the initial data set to obtain a data set to be detected, wherein the data set to be detected comprises content data corresponding to a target object and historical behavior data corresponding to a target user;
generating a first label corresponding to all the objects according to the content data, and generating a second label corresponding to the target user according to the historical behavior data;
respectively calculating the similarity between the corresponding data of all the first labels;
recalling the candidate set corresponding to the target user according to the similarity and the second label;
and sequencing the candidate set to obtain a target set.
According to a specific implementation manner of the embodiment of the present disclosure, the step of performing a preprocessing operation on the initial data set to obtain a data set to be detected includes:
aggregating data which are depended on by a preset algorithm in the initial data;
converting all aggregated data into structured data;
the structured data is classified to form the content data and the historical behavior data.
According to a specific implementation manner of the embodiment of the present disclosure, the step of converting all aggregated data into structured data includes:
converting all aggregated data into numerical data through feature processing operation;
and taking all the numerical value class data as the structured data.
According to a specific implementation manner of the embodiment of the present disclosure, the step of generating the first tag corresponding to all the subjects according to the content data includes:
extracting category features from all the content data;
and generating first labels corresponding to all the objects according to the category characteristics.
According to a specific implementation manner of the embodiment of the present disclosure, the step of generating the second tag corresponding to the target user according to the historical behavior data includes:
filtering interference data in the historical behavior data to obtain effective behavior data, wherein the interference data comprises exposure historical data and negative feedback data;
and generating a second label corresponding to the target user according to the effective behavior data.
According to a specific implementation manner of the embodiment of the present disclosure, the step of calculating the similarity between the data corresponding to all the first tags includes:
converting all the data corresponding to the first label into vector values;
and respectively calculating distance values between all vector values by using a cosine formula and taking the distance values as the similarity.
According to a specific implementation manner of the embodiment of the present disclosure, the step of ranking the candidate sets to obtain the target set includes:
inputting the data in the candidate set into a convolutional neural network to obtain a sequencing model;
and generating the target set according to the sequencing model.
In a second aspect, an embodiment of the present disclosure provides a multi-recall recommendation apparatus, including:
the system comprises a preprocessing module, a data acquisition module and a data processing module, wherein the preprocessing module is used for preprocessing an initial data set to obtain a data set to be detected, and the data set to be detected comprises content data corresponding to a target object and historical behavior data corresponding to a target user;
the generating module is used for generating first labels corresponding to all objects according to the content data and generating second labels corresponding to the target users according to the historical behavior data;
the calculation module is used for calculating the similarity between the data corresponding to all the first labels;
the recalling module is used for recalling the candidate set corresponding to the target user according to the similarity and the second label;
and the sorting module is used for sorting the candidate set to obtain a target set.
In a third aspect, an embodiment of the present disclosure further provides an electronic device, where the electronic device includes:
at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the multi-recall recommendation method of the first aspect or any implementation of the first aspect.
In a fourth aspect, the disclosed embodiments also provide a non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform the multi-recall recommendation method of the first aspect or any implementation manner of the first aspect.
In a fifth aspect, the present disclosure also provides a computer program product including a computer program stored on a non-transitory computer readable storage medium, the computer program including program instructions that, when executed by a computer, cause the computer to perform the multi-recall recommendation method of the first aspect or any implementation manner of the first aspect.
The multi-recall recommendation scheme in the embodiments of the present disclosure includes: preprocessing the initial data set to obtain a data set to be detected, wherein the data set to be detected comprises content data corresponding to a target object and historical behavior data corresponding to a target user; generating a first label corresponding to all the objects according to the content data, and generating a second label corresponding to the target user according to the historical behavior data; respectively calculating the similarity between the corresponding data of all the first labels; recalling the candidate set corresponding to the target user according to the similarity and the second label; and sequencing the candidate set to obtain a target set. According to the scheme, different types of initial data are processed to obtain classified data, corresponding labels are generated, the similarity between the data corresponding to the first labels is calculated, the candidate sets are recalled in combination with the second labels, the candidate sets are ranked to obtain the target set, and the calculation efficiency and recommendation precision of multi-path recall recommendation are improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings needed to be used in the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present disclosure, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a schematic flow chart illustrating a multi-recall recommendation method according to an embodiment of the present disclosure;
FIG. 2 is a partial flow diagram of a multi-recall recommendation method according to an embodiment of the present disclosure;
FIG. 3 is a schematic structural diagram of a multi-recall recommendation device according to an embodiment of the present disclosure;
fig. 4 is a schematic view of an electronic device provided in an embodiment of the present disclosure.
Detailed Description
The embodiments of the present disclosure are described in detail below with reference to the accompanying drawings.
The embodiments of the present disclosure are described below with specific examples, and other advantages and effects of the present disclosure will be readily apparent to those skilled in the art from the disclosure in the specification. It is to be understood that the described embodiments are merely illustrative of some, and not restrictive, of the embodiments of the disclosure. The disclosure may be embodied or carried out in various other specific embodiments, and various modifications and changes may be made in the details within the description without departing from the spirit of the disclosure. It is to be noted that the features in the following embodiments and examples may be combined with each other without conflict. All other embodiments, which can be derived by a person skilled in the art from the embodiments disclosed herein without making any creative effort, shall fall within the protection scope of the present disclosure.
It is noted that various aspects of the embodiments are described below within the scope of the appended claims. It should be apparent that the aspects described herein may be embodied in a wide variety of forms and that any specific structure and/or function described herein is merely illustrative. Based on the disclosure, one skilled in the art should appreciate that one aspect described herein may be implemented independently of any other aspects and that two or more of these aspects may be combined in various ways. For example, an apparatus may be implemented and/or a method practiced using any number of the aspects set forth herein. Additionally, such an apparatus may be implemented and/or such a method may be practiced using other structure and/or functionality in addition to one or more of the aspects set forth herein.
It should be noted that the drawings provided in the following embodiments are only for illustrating the basic idea of the present disclosure, and the drawings only show the components related to the present disclosure rather than the number, shape and size of the components in actual implementation, and the type, amount and ratio of the components in actual implementation may be changed arbitrarily, and the layout of the components may be more complicated.
In addition, in the following description, specific details are provided to facilitate a thorough understanding of the examples. However, it will be understood by those skilled in the art that the aspects may be practiced without these specific details.
At present, the internet can provide massive information for users, and an intelligent recommendation system can help the users to find interesting information quickly. The recommendation algorithm is characterized in that the association relationship between people and articles is established, a core is formed by surrounding data, algorithm and a system, massive data information is applied to a corresponding recall strategy and a corresponding sorting strategy by utilizing the algorithm of the recommendation system, and the personalized recommendation process provided for a user is realized. However, most of the existing recall recommendation methods deal with original data of numerical types, so that little and little data of texts, audios and pictures are involved, recall recommendation is performed on single classified data, and the recommendation efficiency and accuracy are poor. The embodiment of the disclosure provides a multi-channel recall recommendation method, which can be applied to the process of generating recommendation data in scenes such as online shopping malls.
Referring to fig. 1, a flow chart of a multi-recall recommendation method provided in the embodiment of the present disclosure is schematically illustrated. As shown in fig. 1, the method mainly comprises the following steps:
s101, preprocessing an initial data set to obtain a data set to be detected, wherein the data set to be detected comprises content data corresponding to a target object and historical behavior data corresponding to a target user;
in specific implementation, the received initial data set may include data with different standards, such as text, audio, or image, and the like, and the initial data set may be subjected to the preprocessing operation in a unified manner, so as to convert different types of data into the data set to be detected with a consistent format, where the data set to be detected may include content data corresponding to the object and historical behavior data corresponding to the target user, and of course, other classifications may also be set according to actual requirements.
S102, generating first labels corresponding to all objects according to the content data, and generating second labels corresponding to the target users according to the historical behavior data;
for example, the first tag generated according to the content data may include a category, a keyword, a quality score, or the like, and the second tag generated according to the historical behavior data may include browsing, collecting, or purchasing, or the like, corresponding to the target user.
S103, respectively calculating the similarity between the corresponding data of all the first labels;
in specific implementation, the similarity between the data corresponding to all the first tags is respectively calculated, so that the data corresponding to all the first tags form association, and subsequent operation is performed conveniently. Specifically, the data corresponding to all the first labels may be subjected to partial bucket division by using locality sensitive hashing, and after the partial bucket division, only the similarity between the data in the bucket is calculated, so as to improve the calculation efficiency.
S104, recalling the candidate set corresponding to the target user according to the similarity and the second label;
in specific implementation, the similarity and the second label are collected, and the data corresponding to the target user is recalled to form the candidate set.
S105, sorting the candidate set to obtain a target set.
After the candidate set is obtained, the data in the candidate set represents the recommended objects corresponding to the target user, the candidate set can be ranked to obtain the target set, and different objects can be recommended to the target user according to the ranking preference in the target set.
According to the multi-path recall recommendation method, the classification data are obtained after the different types of initial data are processed, the corresponding labels are generated, the similarity between the data corresponding to the first labels is calculated and the candidate set is recalled by combining the second labels, then the candidate set is ranked to obtain the target set, and the calculation efficiency and the recommendation precision of the multi-path recall recommendation are improved.
On the basis of the foregoing embodiment, in step S101, performing a preprocessing operation on the initial data set to obtain a data set to be detected includes:
aggregating data which are depended on by a preset algorithm in the initial data;
in specific implementation, considering that the types of data corresponding to the subject matter may have diversity, the scattered data needs to be aggregated to facilitate subsequent uniform processing, and for the recommendation system, the data sources that depend on are diverse, so that it is very necessary to aggregate all the data that the algorithms depend on. User behavior data is generally uploaded to a log collection web service (such as a Nginx server) through a http protocol at a client site, and fault tolerance and expansibility of log collection can be increased through domain name splitting or LB load balancing service in the middle. The log is generally processed by two streams, i.e. offline and real-time, offline enters the data warehouse through ETL, and real-time stream is processed by a real-time processing program such as Spark Streaming through a message queue such as Kafka or enters a real-time storage such as HBase and elastic search for later service query.
Converting all aggregated data into structured data;
after the initial data set formed by multiple types of data is obtained, operations such as data cleaning, format conversion, missing value filling, elimination repetition and the like can be performed on all data in the initial data set, and finally a piece of structured data with uniform format, high degree of structuralization, high data quality and good compatibility is obtained for feature engineering processing of a recommendation algorithm.
The structured data is classified to form the content data and the historical behavior data.
After the structured data is obtained, the structured data may be classified to form the content data and the historical behavior data, and of course, other classifications may also be generated as needed.
Further, the step of converting all aggregated data into structured data includes:
converting all aggregated data into numerical data through feature processing operation;
and taking all the numerical value class data as the structured data.
In specific implementation, the processes of removing dirty data, checking the validity of the data, removing invalid fields, checking field formats and the like can be realized. The format conversion is a process of converting the same type of data from different sources into the same format according to the definition and the requirement of a recommendation algorithm on the data so as to unify and standardize the data. Due to the problems existing in log dotting or various problems existing in the data collection process (for example, a user generally cannot fill in complete data), field value missing exists certainly in a real service scene, missing value filling is performed according to an average or mode or learning filling is performed by using an algorithm, and then the numerical data is used as the structured data.
On the basis of the foregoing embodiment, as shown in fig. 2, the step S102 of generating the first tag corresponding to all the objects according to the content data includes:
s201, extracting category characteristics from all the content data;
in a specific implementation, the category features may be extracted from the content data, where the content data may include discrete data such as registered capital and provincial regions, text data such as enterprise information and user follow-up records, and audio data such as customer call records, and different operations are required to be performed on different types of data, so as to extract corresponding category features.
For example, discrete features are a very common class of features, and the user attribute data and the object attribute data include a large number of category features, such as gender, academic calendar, type of video, tag, director, country, and the like. For discrete features, the features are encoded in several ways as follows. One-hot coding is generally used for class features, if a certain class feature has k classes, we fix the k classes with an order relationship (there is no matter what order relationship is convenient, and it is only convenient to confirm where a certain class is located), we map each value into a k-dimensional vector, where the component of the value is 1, and the other components are 0. This approach can become very large when the number of classes is large. In this case, dimensionality reduction is performed using a method such as PCA.
For the category feature of tags, each video may have multiple tags, and then one-hot encoding is expanded to n-hot encoding, that is, the video has 1 component corresponding to all tags contained in the video, and the others are 0.
Secondly, for some category characteristics with extremely large values, such as enterprise labels, the hash codes are tens of thousands, the characteristic matrix obtained by using one-hot codes is very sparse, and if characteristic intersection is carried out, the characteristic dimension can be increased explosively. The purpose of the feature hash is to compress an original high-dimensional feature vector into a lower-dimensional feature vector, and the expression capability of the original feature is not lost as much as possible. And the characteristic dimension is reduced, the algorithm training and prediction can be accelerated, and the memory consumption is reduced.
For the text data, an NLP project can be used for converting the text data into dense vector representation, and for the audio data, a third-party interface can be called firstly to convert voice into characters, and then the text data processing operation is carried out. Of course, the numerical data which can be directly used can also be subjected to discretization, the sparse vector obtained after discretization has higher operation speed, and the calculation result is easy to store. The features after discretization are also more robust to outliers.
S202, generating first labels corresponding to all the objects according to the category characteristics.
After the category features are obtained, first tags corresponding to all the subject matters can be generated according to the category features.
Further, the step of generating the second tag corresponding to the target user according to the historical behavior data in step S102 includes:
filtering interference data in the historical behavior data to obtain effective behavior data, wherein the interference data comprises exposure historical data and negative feedback data;
in specific implementation, considering that the behavior of the target user may include negative feedback data of the target object, such as adding a blacklist or exposure history data, the exposure history data and the negative feedback data may be used as the interference data, and after filtering the interference data in the history behavior data, the effective behavior data is obtained, so as to perform the next operation flow.
And generating a second label corresponding to the target user according to the effective behavior data.
And after the effective behavior data is obtained, generating a second label corresponding to the target user according to the effective behavior data.
On the basis of the foregoing embodiment, the step S103 of calculating the similarity between all the data corresponding to the first tag includes:
converting all the data corresponding to the first label into vector values;
in specific implementation, data corresponding to all the first tags may be converted into space vectors, and different data correspond to different vector values.
And respectively calculating distance values between all vector values by using a cosine formula and taking the distance values as the similarity.
The distance between all the vector values can be calculated by using a cosine formula, and the distance value is taken as the similarity to represent the relevance between different data.
Optionally, in step S105, the sorting the candidate sets to obtain a target set includes:
inputting the data in the candidate set into a convolutional neural network to obtain a sequencing model;
in specific implementation, all data in the candidate set may be input to a convolutional neural network for training, so as to obtain the ranking model.
In a specific implementation manner, the ranking model may be a Deep FM learning model, and the Deep FM learning model has the capability of automatically learning cross features, so that the work of shallow artificial feature engineering in the original Wide & Deep model is avoided. The original input features are shared. The original features of the Deep FM model are used as the common input of the FM and Deep model parts, and the accuracy and consistency of the model features are ensured. Where the deep part is a simple feed forward network. In the input feature part, because the original feature vectors are mostly high-latitude, highly sparse, continuous and category-mixed domain features, in order to better exert the capability of learning high-order features of the DNN model, the original sparse representation features can be mapped into dense feature vectors by designing corresponding sub-network structures. When the sub-network structure is designed, different field characteristic lengths need to be guaranteed, but vectors output by the sub-network need to have the same dimensionality, and the sub-network output vectors are obtained by initializing with hidden characteristic vectors V of an FM model as network weights.
And generating the target set according to the sequencing model.
After the ranking model is obtained, the target set may be generated according to the ranking model, or the ranking model may be input after new data of the target user is obtained, so as to obtain a target set corresponding to the target user.
In correspondence with the above method embodiment, referring to fig. 3, the disclosed embodiment further provides a multi-recall recommendation device 30, including:
the preprocessing module 301 is configured to perform preprocessing operation on an initial data set to obtain a data set to be detected, where the data set to be detected includes content data corresponding to a target object and historical behavior data corresponding to a target user;
a generating module 302, configured to generate a first tag corresponding to all objects according to the content data, and generate a second tag corresponding to the target user according to the historical behavior data;
a calculating module 303, configured to calculate similarities between data corresponding to all the first tags respectively;
a recalling module 304, configured to recall the candidate set corresponding to the target user according to the similarity and the second tag;
a sorting module 305, configured to sort the candidate set to obtain a target set.
The apparatus shown in fig. 3 may correspondingly execute the content in the above method embodiment, and details of the part not described in detail in this embodiment refer to the content described in the above method embodiment, which is not described again here.
Referring to fig. 4, an embodiment of the present disclosure also provides an electronic device 40, including:
at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the multi-recall recommendation method of the method embodiments described above.
The disclosed embodiments also provide a non-transitory computer-readable storage medium storing computer instructions for causing the computer to perform the multi-recall recommendation method in the aforementioned method embodiments.
The disclosed embodiments also provide a computer program product comprising a computer program stored on a non-transitory computer readable storage medium, the computer program comprising program instructions that, when executed by a computer, cause the computer to perform the multi-recall recommendation method of the aforementioned method embodiments.
Referring now to FIG. 4, a block diagram of an electronic device 40 suitable for use in implementing embodiments of the present disclosure is shown. The electronic devices in the embodiments of the present disclosure may include, but are not limited to, mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., car navigation terminals), and the like, and fixed terminals such as digital TVs, desktop computers, and the like. The electronic device shown in fig. 4 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiments of the present disclosure.
As shown in fig. 4, the electronic device 40 may include a processing means (e.g., a central processing unit, a graphics processor, etc.) 401 that may perform various appropriate actions and processes in accordance with a program stored in a Read Only Memory (ROM)402 or a program loaded from a storage means 408 into a Random Access Memory (RAM) 403. In the RAM 403, various programs and data necessary for the operation of the electronic apparatus 40 are also stored. The processing device 401, the ROM 402, and the RAM 403 are connected to each other via a bus 404. An input/output (I/O) interface 405 is also connected to bus 404.
Generally, the following devices may be connected to the I/O interface 405: input devices 406 including, for example, a touch screen, touch pad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; an output device 407 including, for example, a Liquid Crystal Display (LCD), a speaker, a vibrator, and the like; storage 408 including, for example, tape, hard disk, etc.; and a communication device 409. The communication device 409 may allow the electronic device 40 to communicate wirelessly or by wire with other devices to exchange data. While the figures illustrate an electronic device 40 having various means, it is to be understood that not all illustrated means are required to be implemented or provided. More or fewer devices may alternatively be implemented or provided.
In particular, according to an embodiment of the present disclosure, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method illustrated in the flow chart. In such an embodiment, the computer program may be downloaded and installed from a network via the communication device 409, or from the storage device 408, or from the ROM 402. The computer program performs the above-described functions defined in the methods of the embodiments of the present disclosure when executed by the processing device 401.
It should be noted that the computer readable medium in the present disclosure can be a computer readable signal medium or a computer readable storage medium or any combination of the two. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples of the computer readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present disclosure, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In contrast, in the present disclosure, a computer readable signal medium may comprise a propagated data signal with computer readable program code embodied therein, either in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: electrical wires, optical cables, RF (radio frequency), etc., or any suitable combination of the foregoing.
The computer readable medium may be embodied in the electronic device; or may exist separately without being assembled into the electronic device.
The computer readable medium carries one or more programs which, when executed by the electronic device, cause the electronic device to perform the steps associated with the method embodiments.
Alternatively, the computer readable medium carries one or more programs which, when executed by the electronic device, enable the electronic device to perform the steps associated with the method embodiments.
Computer program code for carrying out operations for aspects of the present disclosure may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C + +, and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider).
The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
The units described in the embodiments of the present disclosure may be implemented by software or hardware.
It should be understood that portions of the present disclosure may be implemented in hardware, software, firmware, or a combination thereof.
The above description is only for the specific embodiments of the present disclosure, but the scope of the present disclosure is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present disclosure should be covered within the scope of the present disclosure. Therefore, the protection scope of the present disclosure shall be subject to the protection scope of the claims.
Claims (10)
1. A multi-recall recommendation method, comprising:
preprocessing the initial data set to obtain a data set to be detected, wherein the data set to be detected comprises content data corresponding to a target object and historical behavior data corresponding to a target user;
generating a first label corresponding to all the objects according to the content data, and generating a second label corresponding to the target user according to the historical behavior data;
respectively calculating the similarity between the corresponding data of all the first labels;
recalling the candidate set corresponding to the target user according to the similarity and the second label;
and sequencing the candidate set to obtain a target set.
2. The method according to claim 1, wherein the step of performing a preprocessing operation on the initial data set to obtain the data set to be detected comprises:
aggregating data which are depended on by a preset algorithm in the initial data;
converting all aggregated data into structured data;
the structured data is classified to form the content data and the historical behavior data.
3. The method of claim 2, wherein the step of converting the aggregated total data into structured data comprises:
converting all aggregated data into numerical data through feature processing operation;
and taking all the numerical value class data as the structured data.
4. The method of claim 2, wherein the step of generating a first tag corresponding to all objects from the content data comprises:
extracting category features from all the content data;
and generating first labels corresponding to all the objects according to the category characteristics.
5. The method according to claim 2, wherein the step of generating the second label corresponding to the target user according to the historical behavior data includes:
filtering interference data in the historical behavior data to obtain effective behavior data, wherein the interference data comprises exposure historical data and negative feedback data;
and generating a second label corresponding to the target user according to the effective behavior data.
6. The method according to claim 1, wherein the step of calculating the similarity between all the first tag corresponding data comprises:
converting all the data corresponding to the first label into vector values;
and respectively calculating distance values between all vector values by using a cosine formula and taking the distance values as the similarity.
7. The method of claim 1, wherein the step of ranking the candidate set to obtain a target set comprises:
inputting the data in the candidate set into a convolutional neural network to obtain a sequencing model;
and generating the target set according to the sequencing model.
8. A multi-recall recommendation device, comprising:
the system comprises a preprocessing module, a data acquisition module and a data processing module, wherein the preprocessing module is used for preprocessing an initial data set to obtain a data set to be detected, and the data set to be detected comprises content data corresponding to a target object and historical behavior data corresponding to a target user;
the generating module is used for generating first labels corresponding to all objects according to the content data and generating second labels corresponding to the target users according to the historical behavior data;
the calculation module is used for calculating the similarity between the data corresponding to all the first labels;
the recalling module is used for recalling the candidate set corresponding to the target user according to the similarity and the second label;
and the sorting module is used for sorting the candidate set to obtain a target set.
9. An electronic device, characterized in that the electronic device comprises:
at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the multi-recall recommendation method of any of preceding claims 1-7.
10. A non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform the multi-recall recommendation method of any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110038094.1A CN112685516A (en) | 2021-01-12 | 2021-01-12 | Multi-channel recall recommendation method and device, electronic equipment and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110038094.1A CN112685516A (en) | 2021-01-12 | 2021-01-12 | Multi-channel recall recommendation method and device, electronic equipment and medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112685516A true CN112685516A (en) | 2021-04-20 |
Family
ID=75457604
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110038094.1A Pending CN112685516A (en) | 2021-01-12 | 2021-01-12 | Multi-channel recall recommendation method and device, electronic equipment and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112685516A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113139834A (en) * | 2021-04-29 | 2021-07-20 | 北京沃东天骏信息技术有限公司 | Information processing method, device, electronic equipment and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109508423A (en) * | 2018-12-14 | 2019-03-22 | 平安科技（深圳）有限公司 | Source of houses recommended method, device, equipment and storage medium based on semantics recognition |
| CN111210258A (en) * | 2019-12-23 | 2020-05-29 | 北京三快在线科技有限公司 | Advertisement putting method and device, electronic equipment and readable storage medium |
| CN111858915A (en) * | 2020-08-07 | 2020-10-30 | 成都理工大学 | Information recommendation method and system based on label similarity |
| CA3081276A1 (en) * | 2019-06-30 | 2020-12-30 | The Toronto-Dominion Bank | Systems and methods for generating and adjusting recommendations provided on a user interface |
-
2021
- 2021-01-12 CN CN202110038094.1A patent/CN112685516A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109508423A (en) * | 2018-12-14 | 2019-03-22 | 平安科技（深圳）有限公司 | Source of houses recommended method, device, equipment and storage medium based on semantics recognition |
| CA3081276A1 (en) * | 2019-06-30 | 2020-12-30 | The Toronto-Dominion Bank | Systems and methods for generating and adjusting recommendations provided on a user interface |
| CN111210258A (en) * | 2019-12-23 | 2020-05-29 | 北京三快在线科技有限公司 | Advertisement putting method and device, electronic equipment and readable storage medium |
| CN111858915A (en) * | 2020-08-07 | 2020-10-30 | 成都理工大学 | Information recommendation method and system based on label similarity |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113139834A (en) * | 2021-04-29 | 2021-07-20 | 北京沃东天骏信息技术有限公司 | Information processing method, device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109492772B (en) | Method and device for generating information | |
| CN110377740B (en) | Emotion polarity analysis method and device, electronic equipment and storage medium | |
| CN107944481B (en) | Method and apparatus for generating information | |
| CN111797210A (en) | Information recommendation method, device and equipment based on user portrait and storage medium | |
| CN112035743B (en) | Data recommendation method and device, computer equipment and storage medium | |
| CN112364860B (en) | Training method and device of character recognition model and electronic equipment | |
| WO2022121801A1 (en) | Information processing method and apparatus, and electronic device | |
| CN110633423B (en) | Target account identification method, device, equipment and storage medium | |
| CN110278447B (en) | Video pushing method and device based on continuous features and electronic equipment | |
| CN113688310B (en) | Content recommendation method, device, equipment and storage medium | |
| CN110457325B (en) | Method and apparatus for outputting information | |
| CN110059172B (en) | Method and device for recommending answers based on natural language understanding | |
| CN113051933B (en) | Model training method, text semantic similarity determination method, device and equipment | |
| EP4083860A1 (en) | Method and apparatus for training item coding model | |
| CN114417169A (en) | Information recommendation optimization method, device, medium, and program product | |
| CN113971400B (en) | Text detection method and device, electronic equipment and storage medium | |
| US11437038B2 (en) | Recognition and restructuring of previously presented materials | |
| CN111382365B (en) | Method and device for outputting information | |
| CN112685516A (en) | Multi-channel recall recommendation method and device, electronic equipment and medium | |
| CN113761174A (en) | Text generation method and device | |
| CN112035256A (en) | Resource allocation method, device, electronic equipment and medium | |
| CN111382262A (en) | Method and apparatus for outputting information | |
| CN116308704A (en) | Product recommendation method, device, electronic equipment, medium and computer program product | |
| CN112801053B (en) | Video data processing method and device | |
| CN114969427A (en) | Singing list generation method and device, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210420 |
|
| RJ01 | Rejection of invention patent application after publication |