CN109684549A - Target data prediction method and device, electronic equipment and computer storage medium - Google Patents
Target data prediction method and device, electronic equipment and computer storage medium Download PDFInfo
- Publication number
- CN109684549A CN109684549A CN201811583663.5A CN201811583663A CN109684549A CN 109684549 A CN109684549 A CN 109684549A CN 201811583663 A CN201811583663 A CN 201811583663A CN 109684549 A CN109684549 A CN 109684549A
- Authority
- CN
- China
- Prior art keywords
- data
- user
- feature data
- fisrt feature
- active user
- 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 48
- 238000012549 training Methods 0.000 claims description 21
- 238000006243 chemical reaction Methods 0.000 abstract description 8
- 230000006399 behavior Effects 0.000 description 24
- 238000010586 diagram Methods 0.000 description 16
- 230000003542 behavioural effect Effects 0.000 description 9
- 230000006870 function Effects 0.000 description 8
- 230000006854 communication Effects 0.000 description 6
- 238000004891 communication Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 5
- 238000004590 computer program Methods 0.000 description 4
- 241001269238 Data Species 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 235000007926 Craterellus fallax Nutrition 0.000 description 2
- 240000007175 Datura inoxia Species 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 238000000151 deposition Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
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
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the disclosure discloses a target data prediction method, a target data prediction device, an electronic device and a computer storage medium, wherein the target data prediction method comprises the following steps: acquiring first characteristic data of a current user, and predicting the category of the current user according to the first characteristic data and a user category prediction model; performing feature filling on the first feature data according to the feature data of the historical user included in the current user category to obtain second feature data of the current user; and predicting the target data according to the second characteristic data. According to the technical scheme, the accuracy of new user data prediction can be improved, the push information matched with the preference of the current user is provided for the current user, the current user can conveniently check or select the push information, the use experience of the user is improved, and the transaction conversion rate is improved.
Description
Technical field
This disclosure relates to technical field of data prediction, and in particular to a kind of target data prediction technique, device, electronic equipment
And computer storage medium.
Background technique
With the development of internet technology, more and more businessmans or service provider by internet platform come for
Family provides service.In order to improve service quality, promoted the usage experience of user, while improving the efficiency of user's operation, Hen Duoping
Platform all when user is using servicing, pushes information relevant to service to user and saves it to facilitate it to check or select and search
The time of rope information.Use several schemes for user's pushed information in the prior art: 1, random push, this scheme execute
It is relatively simple, but the information pushed is not often matched that with user preferences, therefore is not only not achieved and it is facilitated to check or select
The purpose selected can also reduce its usage experience；2, it is pushed based on preset rules, this scheme is strong for the dependence of rule, difficult
To cover whole users, and the adjustment and variation of rule also will increase the difficulty of maintenance；3, it is gone through based on what platform was traded
The characteristic of history user trains information select probability prediction model, the select probability that right rear line push prediction obtains compared with
High information, for the program compares first two scheme, push effect is preferable, but since information select probability prediction model is made
Training data is to have occurred and that the user data of transaction in platform, and the model obtained using the training of these data is used old
The forecasting accuracy at family is higher, but lower for the new user in predicting accuracy for having not occurred transaction, and transaction conversion ratio is lower.
Summary of the invention
The embodiment of the present disclosure provides a kind of target data prediction technique, device, electronic equipment and computer storage medium.
In a first aspect, providing a kind of target data prediction technique in the embodiment of the present disclosure.
Specifically, the target data prediction technique, comprising:
Obtain the fisrt feature data of active user；
The classification of the active user is predicted according to the fisrt feature data and class of subscriber prediction model；
The characteristic for the historical user for being included according to active user's classification for the fisrt feature data into
The filling of row feature, obtains the second feature data of the active user；
The target data of active user is predicted according to the second feature data.
With reference to first aspect, the disclosure is described according to the active user in the first implementation of first aspect
The characteristic for the historical user that classification is included carries out feature filling for the fisrt feature data, obtains the current use
The second feature data at family, comprising:
Determine the feature of missing data in the fisrt feature data；
Historical use data is obtained to be filled the feature of missing data in the fisrt feature data.
With reference to first aspect with the first implementation of first aspect, second in first aspect of the embodiment of the present invention
In implementation, the acquisition historical use data is filled the feature of missing data in the fisrt feature data, packet
It includes:
Calculate the mean value of the missing data for the historical user that active user's classification is included；
Obtained missing data mean value is added in the fisrt feature data, the second feature of the active user is obtained
Data.
With reference to first aspect, second of implementation of the first implementation of first aspect and first aspect, this hair
Bright embodiment before the fisrt feature data for obtaining active user, is also wrapped in the third implementation of first aspect
It includes:
The fisrt feature data and its classification information of the historical user in default historical time section are obtained, and are gone through according to described
The fisrt feature data of history user and the training of corresponding classification information obtain the class of subscriber prediction model.
With reference to first aspect, the first implementation of first aspect, first aspect second of implementation and first
The third implementation of aspect, for the embodiment of the present invention in the 4th kind of implementation of first aspect, the class of subscriber is pre-
Survey model is maximum entropy model.
With reference to first aspect, the first implementation, second of implementation of first aspect, first party of first aspect
The third implementation in face and the 4th kind of implementation of first aspect, five kind reality of the embodiment of the present invention in first aspect
It is described that the target data of active user is predicted according to the second feature data in existing mode, comprising:
Determine target data prediction model；
According to the second feature data, using the target data prediction model for active user target data into
Row prediction.
With reference to first aspect, the first implementation, second of implementation of first aspect, first party of first aspect
The 5th kind of implementation of the third implementation in face, the 4th kind of implementation of first aspect and first aspect, the present invention
For embodiment in the 6th kind of implementation of first aspect, the target data prediction model is LR model.
With reference to first aspect, the first implementation, second of implementation of first aspect, first party of first aspect
The third implementation in face, the 4th kind of implementation of first aspect, first aspect the 5th kind of implementation and first party
The 6th kind of implementation in face, the embodiment of the present invention is in the 7th kind of implementation of first aspect, further includes:
Predetermined registration operation is executed according to the target data that prediction obtains.
Second aspect provides a kind of target data prediction meanss in the embodiment of the present disclosure.
Specifically, the target data prediction meanss, comprising:
First prediction module is configured as obtaining the fisrt feature data of active user；According to the fisrt feature data
The classification of the active user is predicted with class of subscriber prediction model；
Module is filled, is configured as the characteristic for the historical user for being included according to active user's classification for institute
It states fisrt feature data and carries out feature filling, obtain the second feature data of the active user；
Second prediction module is configured as being carried out according to target data of the second feature data for active user pre-
It surveys.
In conjunction with second aspect, the embodiment of the present invention is in the first implementation of second aspect, the filling module packet
It includes:
First determines submodule, is configured to determine that the feature of missing data in the fisrt feature data；
Submodule is filled, is configured as obtaining historical use data to the feature of missing data in the fisrt feature data
It is filled.
In conjunction with the first of second aspect and second aspect implementation, second in second aspect of the embodiment of the present invention
In implementation, the filling submodule includes:
Computational submodule is configured as calculating the equal of the missing data for the historical user that active user's classification is included
Value；
Submodule is added, is configured as the missing data mean value that will be obtained and is added in the fisrt feature data, obtain institute
State the second feature data of active user.
In conjunction with the first implementation of second aspect, second aspect and second of implementation of second aspect, this hair
Bright embodiment is in the third implementation of second aspect, further includes:
Training module is configured as obtaining the fisrt feature data and its classification of the historical user in default historical time section
Information, and the class of subscriber is obtained according to the fisrt feature data of the historical user and the training of corresponding classification information and is predicted
Model.
In conjunction with the first implementation of second aspect, second aspect, second of implementation and second of second aspect
The third implementation of aspect, for the embodiment of the present invention in the 4th kind of implementation of second aspect, the class of subscriber is pre-
Survey model is maximum entropy model.
The first implementation, second of implementation of second aspect, second party in conjunction with second aspect, second aspect
The third implementation in face and the 4th kind of implementation of second aspect, five kind reality of the embodiment of the present invention in second aspect
In existing mode, second prediction module includes:
Second determines submodule, is configured to determine that target data prediction model；
Predict submodule, be configured as according to the second feature data, using the target data prediction model for
The target data of active user is predicted.
The first implementation, second of implementation of second aspect, second party in conjunction with second aspect, second aspect
The 5th kind of implementation of the third implementation in face, the 4th kind of implementation of second aspect and second aspect, the present invention
For embodiment in the 6th kind of implementation of second aspect, the target data prediction model is LR model.
The first implementation, second of implementation of second aspect, second party in conjunction with second aspect, second aspect
The third implementation in face, the 4th kind of implementation of second aspect, second aspect the 5th kind of implementation and second party
The 6th kind of implementation in face, the embodiment of the present invention is in the 7th kind of implementation of second aspect, further includes:
Execution module is configured as executing predetermined registration operation according to the target data that prediction obtains.
The third aspect, the embodiment of the present disclosure provide a kind of electronic equipment, including memory and processor, the memory
It is executed in above-mentioned first aspect based on target data prediction technique by storing one or more support target data prediction meanss
Calculation machine instruction, the processor is configured to for executing the computer instruction stored in the memory.The target data
Prediction meanss can also include communication interface, for target data prediction meanss and other equipment or communication.
Fourth aspect, the embodiment of the present disclosure provides a kind of computer readable storage medium, pre- for storing target data
Computer instruction used in device is surveyed, it includes be target data for executing target data prediction technique in above-mentioned first aspect
Computer instruction involved in prediction meanss.
The technical solution that the embodiment of the present disclosure provides can include the following benefits:
The characteristic of the comprehensive usage history user of above-mentioned technical proposal and active user carry out the prediction of target data, this
The characteristic information that sample can either embody active user can safeguard the standard of target data prediction by the characteristic information of historical user again
True property.The technical solution can be improved the accuracy of new user data prediction, provides to like with it for active user and matches
Pushed information facilitates it to check or select, to promote the usage experience of user, improves transaction conversion ratio.
It should be understood that above general description and following detailed description be only it is exemplary and explanatory, not
The disclosure can be limited.
Detailed description of the invention
In conjunction with attached drawing, by the detailed description of following non-limiting embodiment, the other feature of the disclosure, purpose and excellent
Point will be apparent.In the accompanying drawings:
Fig. 1 shows the flow chart of the target data prediction technique according to one embodiment of the disclosure；
Fig. 2 shows the flow charts of the step S101 of the target data prediction technique of embodiment according to Fig. 1；
Fig. 3 shows the flow chart of the step S102 of the target data prediction technique of embodiment according to Fig. 1；
Fig. 4 shows the flow chart of the step S302 of the target data prediction technique of embodiment according to Fig.3,；
Fig. 5 shows the flow chart of the target data prediction technique according to another embodiment of the disclosure；
Fig. 6 shows the flow chart of the step S103 of the target data prediction technique of embodiment according to Fig. 1；
Fig. 7 shows the flow chart of the target data prediction technique according to disclosure a further embodiment；
Fig. 8 shows the structural block diagram of the target data prediction meanss according to one embodiment of the disclosure；
Fig. 9 shows the structural frames of the first prediction module 801 of the target data prediction meanss of embodiment according to Fig.8,
Figure；
Figure 10 shows the structural block diagram of the filling module 802 of the target data prediction meanss of embodiment according to Fig.8,；
Figure 11 shows the structure of the filling submodule 1002 of the target data prediction meanss of embodiment according to Fig.10,
Block diagram；
Figure 12 shows the structural block diagram of the target data prediction meanss according to another embodiment of the disclosure；
Figure 13 shows the structure of the second prediction module 803 of the target data prediction meanss of embodiment according to Fig.8,
Block diagram；
Figure 14 shows the structural block diagram of the target data prediction meanss according to disclosure a further embodiment；
Figure 15 shows the structural block diagram of the electronic equipment according to one embodiment of the disclosure；
Figure 16 is adapted for the computer system for realizing the target data prediction technique according to one embodiment of the disclosure
Structural schematic diagram.
Specific embodiment
Hereinafter, the illustrative embodiments of the disclosure will be described in detail with reference to the attached drawings, so that those skilled in the art can
Easily realize them.In addition, for the sake of clarity, the portion unrelated with description illustrative embodiments is omitted in the accompanying drawings
Point.
In the disclosure, it should be appreciated that the term of " comprising " or " having " etc. is intended to refer to disclosed in this specification
Feature, number, step, behavior, the presence of component, part or combinations thereof, and be not intended to exclude other one or more features,
A possibility that number, step, behavior, component, part or combinations thereof exist or are added.
It also should be noted that in the absence of conflict, the feature in embodiment and embodiment in the disclosure
It can be combined with each other.The disclosure is described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.
The characteristic of overall evaluation of a technical project usage history user and active user that the embodiment of the present disclosure provides carry out mesh
The prediction for marking data, the characteristic information that can either embody active user in this way can safeguard mesh by the characteristic information of historical user again
Mark the accuracy of data prediction.The technical solution can be improved the accuracy of new user data prediction, provided for active user and
It likes the pushed information to match, it is facilitated to check or select, to promote the usage experience of user, improves transaction conversion
Rate.
Fig. 1 shows the flow chart of the target data prediction technique according to one embodiment of the disclosure.As shown in Figure 1, described
Target data prediction technique includes the following steps S101-S103:
In step s101, the fisrt feature data of active user are obtained；According to the fisrt feature data and user class
Other prediction model predicts the classification of the active user；
In step s 102, the characteristic for the historical user for being included according to active user's classification is for described
One characteristic carries out feature filling, obtains the second feature data of the active user；
In step s 103, the target data of active user is predicted according to the second feature data.
Mentioned above, with the development of internet technology, more and more businessmans or service provider pass through internet
Platform for user provides service.In order to improve service quality, promoted the usage experience of user, while improving the effect of user's operation
Rate, many platforms all when user is using servicing, push information relevant to service to user, to facilitate it to check or select
It selects, saves its time for searching for information.But scheme in the prior art or the information of push are not matched that with user preferences, or
Person user's coverage is lower, then for had not occurred transaction new user in predicting accuracy it is low, transaction conversion ratio it is low.
In view of drawbacks described above, in this embodiment, a kind of target data prediction technique is proposed, this method is comprehensive to be used
The characteristic of historical user and active user carry out the prediction of target data, can either embody the feature letter of active user in this way
Breath can safeguard the accuracy of target data prediction by the characteristic information of historical user again.The technical solution can be improved new user
The accuracy of data prediction, provides the pushed information liked with it and matched for active user, it is facilitated to check or select, from
And the usage experience of user is promoted, improve transaction conversion ratio.
In an optional implementation of the present embodiment, the active user refer to currently need for itself or
The new user that its relevant target data of behavior that next may occur is predicted, the historical user refer to once existing
A certain platform is excessively single up and down, clicked some link or occurred transaction etc. implemented to preset the user of vaild act.
In an optional implementation of the present embodiment, the fisrt feature data are referred to from current-user data
The characteristic that can be extracted, for example, attributive character such as the age of active user, gender, occupation etc..Usual situation
Under, for active user as new user, the data volume of related data is less than the data volume of historical user, because historical user is
Relevant behavior operation was occurred in the platform, descended it is single, clicked some link or transaction, corresponding behavior occurred
Operation data is also recorded, it may be said that behavior occurs and operates more users, corresponding user data is also richer
Richness, for example, can also extract to obtain other than the attributive character such as above-mentioned age, gender, occupation, in historical use data objective single
Behaviors operating characteristics or other features such as valence, preferential susceptibility, chain degree of liking.
In an optional implementation of the present embodiment, the target data refers to use that needs are predicted and current
Next the next relevant object of behavior or content-data that family may occur, such as active user are possible to the quotient of selection
Family or service, or next it is possible to the link that can be clicked or information etc..
In an optional implementation of the present embodiment, the class of subscriber prediction model is described current for predicting
The model of user's owning user classification, and the class of subscriber prediction model is obtained based on historical use data training.?
In the implementation, it is believed that the behavioral characteristic for belonging to same category of user has certain similitude.Wherein, the user class
It is not the classification classified according to different user feature for user, for example, can be incited somebody to action according to different user characteristics
User is divided into tri- classifications of A, B, C, and different classes of user characteristics have the difference in apparent numberical range, with age characteristics
For, 18-24 years old user can be divided into A class, 24-40 years old user is divided into B class, 40 years old or more user is divided
For C class, above-mentioned classifying rules is also referred to as following characteristics classification function:
Certainly, when classifying to user, can also be classified according to other features, or comprehensive multiple features one
And classify, those skilled in the art can be configured according to the needs of practical application, and the disclosure is not especially limited it.
In an optional implementation of the present embodiment, the second feature data are more than the fisrt feature data
It is to utilize the spy for the historical user that active user's classification is included based on the fisrt feature data for complete characteristic
Sign data carry out the fisrt feature data characteristic obtained after feature filling, the comparatively complete characteristic
According to the subsequent data basis that will be predicted as target data, to improve the accuracy of target data prediction.
In an optional implementation of the present embodiment, as shown in Fig. 2, the step S101, i.e. acquisition active user
Fisrt feature data；The classification of the active user is predicted according to the fisrt feature data and class of subscriber prediction model
Step, including step S201-S202:
In step s 201, the fisrt feature data of active user are obtained；
In step S202, the fisrt feature data are input to the class of subscriber prediction model, obtain described work as
The class prediction result of preceding user.
In view of the characteristic of new user is usually fewer, therefore in order to be provided more in the prediction of succeeding target data
The data basis of horn of plenty is predicted the classification of the new user, so as to subsequent according to prediction in this embodiment
Class of subscriber its characteristic is supplemented.Specifically, the fisrt feature data of active user are obtained first；Then will
The fisrt feature data are input to the class of subscriber prediction model, obtain the class prediction result of the active user.
Wherein, the class of subscriber prediction model is obtained according to relatively complete historical use data training.
In an optional implementation of the present embodiment, as shown in figure 3, the step S102, i.e., according to described current
The characteristic for the historical user that class of subscriber is included carries out feature filling for the fisrt feature data, obtains described work as
The step of second feature data of preceding user, including step S301-S302:
In step S301, the feature of missing data in the fisrt feature data is determined；
In step s 302, historical use data is obtained to fill out the feature of missing data in the fisrt feature data
It fills.
In order to which fisrt feature data relatively small number of for the current-user data amount are supplemented, so that fisrt feature
Data are also relatively complete, provide data basis abundant for the prediction of succeeding target data, in this embodiment, using current
The characteristic for the historical user that class of subscriber is included carries out feature filling for the fisrt feature data.Specifically, first
First determine the feature of missing data in the fisrt feature data, wherein the feature of the missing data can be according to the history
Characteristic common to user determines, for example, if characteristic common to the historical user include attributive character and
Behavioural characteristic then the attributive character and behavioural characteristic is relatively determining compared with fisrt feature data can obtain the fisrt feature
The feature of shortage of data data, for example, the feature of missing is exactly to go if the fisrt feature data only include attributive character
It is characterized；Then it obtains historical use data to be filled the feature of missing data in the fisrt feature data, obtains institute
The second feature data of active user are stated, such as, it is assumed that active user's generic is A class, then it is assumed that the current use
There are certain similitudes for the behavior of historical user in the behavior and A class at family, therefore, take the behavioural characteristic of historical user in A class
Feature filling is carried out for the fisrt feature data, obtains the relatively complete second feature data of the active user.
In an optional implementation of the present embodiment, as shown in figure 4, the step S302, i.e. acquisition historical user
The step of data are filled the feature of missing data in the fisrt feature data, including step S401-S402:
In step S401, the mean value of the missing data for the historical user that active user's classification is included is calculated；
In step S402, obtained missing data mean value is added in the fisrt feature data, is obtained described current
The second feature data of user.
In this embodiment, when carrying out feature filling for the fisrt feature data using the missing data,
The mean value of the missing data of the included historical user of active user's classification can be added as feature supplementary data to described
In fisrt feature data, the relatively complete second feature data of the active user are obtained.
For example, if determining the missing data for behaviors such as visitor's unit price, preferential susceptibility, chain degree of liking through above-mentioned analysis
Operation data, and classification belonging to the active user is A class, then the visitor of all historical users in A class can be obtained
Monovalent, preferential susceptibility, it is chain like degree evidence, and it is the monovalent mean value of the visitor for calculating separately all historical users in A class, preferential quick
Sensitivity mean value and chain degree of liking mean value, are supplemented in the fisrt feature data of the active user, are worked as described
The behaviors operation datas such as visitor's unit price of preceding user, preferential susceptibility, chain degree of liking, thus form one and both belong to comprising user
Property feature, and include the second feature data of behavior operating characteristics.
Certainly, the mode of other features filling can also be used, the present invention does not make the specific implementation that feature is filled
Specific to limit, those skilled in the art can according to the needs of practical application and the characteristics of characteristic selects.
In an optional implementation of the present embodiment, the step S101 obtains the fisrt feature of active user
Data；Before the step of predicting the classification of the active user according to the fisrt feature data and class of subscriber prediction model,
Further include the steps that being trained the class of subscriber prediction model, i.e., as shown in figure 5, the method includes the steps
S501-S504:
In step S501, the fisrt feature data and its classification letter of the historical user in default historical time section are obtained
Breath, and the class of subscriber is obtained according to the fisrt feature data of the historical user and the training of corresponding classification information and predicts mould
Type；
In step S502, the fisrt feature data of active user are obtained；According to the fisrt feature data and user class
Other prediction model predicts the classification of the active user；
In step S503, the characteristic for the historical user for being included according to active user's classification is for described
One characteristic carries out feature filling, obtains the second feature data of the active user；
In step S504, the target data of active user is predicted according to the second feature data.
In this embodiment, when being trained for the class of subscriber prediction model, default history is obtained first
The fisrt feature data and its classification information of historical user in period, then according to the fisrt feature number of the historical user
The class of subscriber prediction model is obtained according to the training of corresponding classification information.
In an optional implementation of the present embodiment, the class of subscriber prediction model may be selected to be maximum entropy prediction
Model, the maximum entropy prediction model principle is to keep the entropy of prediction result maximum, i.e., so that the prediction model is estimated with unbiased
Meter.It is of course also possible to select other effective prediction models, those skilled in the art can determine suitable according to the needs of practical application
Prediction model, the present invention are not especially limited it.
In an optional implementation of the present embodiment, as shown in fig. 6, the step S103, i.e., according to described second
The step of characteristic predicts the target data of active user, including step S601-S602:
In step s 601, target data prediction model is determined；
In step S602, according to the second feature data, using the target data prediction model for currently using
The target data at family is predicted.
In this embodiment, the target data of active user is predicted according to the second feature data
When, first according to the target data prediction model for determining needs the characteristics of the second feature data and target data；Then again
According to the second feature data, the target data of active user is predicted using the target data prediction model.
Wherein, the target data prediction model is chosen as the models such as LR model, naturally it is also possible to select other models, the present invention is to it
It is not especially limited.
In an optional implementation of the present embodiment, in the step S602 according to the second feature data, benefit
When being predicted with the target data prediction model for the target data of active user, the prediction models such as LR model may be selected
The target data is predicted, it is of course also possible to select other prediction models.Those skilled in the art can be according to reality
The needs of application select suitable prediction model, and the present invention is not especially limited it.
In an optional implementation of the present embodiment, the method also includes the target datas obtained according to prediction to hold
The step of row predetermined registration operation, i.e., as shown in fig. 7, the method includes the steps S701-S704:
In step s 701, the fisrt feature data of active user are obtained；According to the fisrt feature data and user class
Other prediction model predicts the classification of the active user；
In step S702, the characteristic for the historical user for being included according to active user's classification is for described
One characteristic carries out feature filling, obtains the second feature data of the active user；
In step S703, the target data of active user is predicted according to the second feature data；
In step S704, predetermined registration operation is executed according to the target data that prediction obtains.
After prediction obtains the target data, predetermined registration operation can be executed according to the target data, wherein described default
Operation includes one of following operation or a variety of: by the target data in the current of current interface or current interface user
Operating area shows, sends or be pushed to described active user etc. for the target data.
Following is embodiment of the present disclosure, can be used for executing embodiments of the present disclosure.
Fig. 8 shows the structural block diagram of the target data prediction meanss according to one embodiment of the disclosure, which can lead to
Cross being implemented in combination with as some or all of of electronic equipment of software, hardware or both.As shown in figure 8, the number of targets
It is predicted that device includes:
First prediction module 801 is configured as obtaining the fisrt feature data of active user；According to the fisrt feature number
According to the classification for predicting the active user with class of subscriber prediction model；
Module 802 is filled, the characteristic pair for the historical user for being included according to active user's classification is configured as
Feature filling is carried out in the fisrt feature data, obtains the second feature data of the active user；
Second prediction module 803, be configured as according to the second feature data for active user target data into
Row prediction.
Mentioned above, with the development of internet technology, more and more businessmans or service provider pass through internet
Platform for user provides service.In order to improve service quality, promoted the usage experience of user, while improving the effect of user's operation
Rate, many platforms all when user is using servicing, push information relevant to service to user, to facilitate it to check or select
It selects, saves its time for searching for information.But scheme in the prior art or the information of push are not matched that with user preferences, or
Person user's coverage is lower, then for had not occurred transaction new user in predicting accuracy it is low, transaction conversion ratio it is low.
In view of drawbacks described above, in this embodiment, a kind of target data prediction meanss are proposed, the device is comprehensive to be used
The characteristic of historical user and active user carry out the prediction of target data, can either embody the feature letter of active user in this way
Breath can safeguard the accuracy of target data prediction by the characteristic information of historical user again.The technical solution can be improved new user
The accuracy of data prediction, provides the pushed information liked with it and matched for active user, it is facilitated to check or select, from
And the usage experience of user is promoted, improve transaction conversion ratio.
In an optional implementation of the present embodiment, the active user refer to currently need for itself or
The new user that its relevant target data of behavior that next may occur is predicted, the historical user refer to once existing
A certain platform is excessively single up and down, clicked some link or occurred transaction etc. implemented to preset the user of vaild act.
In an optional implementation of the present embodiment, the fisrt feature data are referred to from current-user data
The characteristic that can be extracted, for example, attributive character such as the age of active user, gender, occupation etc..Usual situation
Under, for active user as new user, the data volume of related data is less than the data volume of historical user, because historical user is
Relevant behavior operation was occurred in the platform, descended it is single, clicked some link or transaction, corresponding behavior occurred
Operation data is also recorded, it may be said that behavior occurs and operates more users, corresponding user data is also richer
Richness, for example, can also extract to obtain other than the attributive character such as above-mentioned age, gender, occupation, in historical use data objective single
Behaviors operating characteristics or other features such as valence, preferential susceptibility, chain degree of liking.
In an optional implementation of the present embodiment, the target data refers to use that needs are predicted and current
Next the next relevant object of behavior or content-data that family may occur, such as active user are possible to the quotient of selection
Family or service, or next it is possible to the link that can be clicked or information etc..
In an optional implementation of the present embodiment, the class of subscriber prediction model is described current for predicting
The model of user's owning user classification, and the class of subscriber prediction model is obtained based on historical use data training.?
In the implementation, it is believed that the behavioral characteristic for belonging to same category of user has certain similitude.Wherein, the user class
It is not the classification classified according to different user feature for user, for example, can be incited somebody to action according to different user characteristics
User is divided into tri- classifications of A, B, C, and different classes of user characteristics have the difference in apparent numberical range, with age characteristics
For, 18-24 years old user can be divided into A class, 24-40 years old user is divided into B class, 40 years old or more user is divided
For C class, above-mentioned classifying rules is also referred to as following characteristics classification function:
Certainly, when classifying to user, can also be classified according to other features, or comprehensive multiple features one
And classify, those skilled in the art can be configured according to the needs of practical application, and the disclosure is not especially limited it.
In an optional implementation of the present embodiment, the second feature data are more than the fisrt feature data
It is to utilize the spy for the historical user that active user's classification is included based on the fisrt feature data for complete characteristic
Sign data carry out the fisrt feature data characteristic obtained after feature filling, the comparatively complete characteristic
According to the subsequent data basis that will be predicted as target data, to improve the accuracy of target data prediction.
In an optional implementation of the present embodiment, as shown in figure 9, first prediction module 801 includes:
Fisrt feature data acquisition submodule 901 is configured as obtaining the fisrt feature data of active user；
Class prediction submodule 902 is configured as the fisrt feature data being input to the class of subscriber prediction mould
Type obtains the class prediction result of the active user.
In view of the characteristic of new user is usually fewer, therefore in order to be provided more in the prediction of succeeding target data
The data basis of horn of plenty is predicted the classification of the new user, so as to subsequent according to prediction in this embodiment
Class of subscriber its characteristic is supplemented.Specifically, fisrt feature data acquisition submodule 901 obtains active user
Fisrt feature data；The fisrt feature data are input to the class of subscriber prediction model by class prediction submodule 902,
Obtain the class prediction result of the active user.
Wherein, the class of subscriber prediction model is obtained according to relatively complete historical use data training.
In an optional implementation of the present embodiment, as shown in Figure 10, the filling module 802 includes:
First determines submodule 1001, is configured to determine that the feature of missing data in the fisrt feature data；
Submodule 1002 is filled, is configured as obtaining historical use data to missing data in the fisrt feature data
Feature is filled.
In order to which fisrt feature data relatively small number of for the current-user data amount are supplemented, so that fisrt feature
Data are also relatively complete, provide data basis abundant for the prediction of succeeding target data, in this embodiment, using current
The characteristic for the historical user that class of subscriber is included carries out feature filling for the fisrt feature data.Specifically,
One determines that submodule 1001 determines the feature of missing data in the fisrt feature data, wherein the feature of the missing data
It can be determined according to characteristic common to the historical user, for example, if characteristic common to the historical user
It, then can be more determining compared with fisrt feature data by the attributive character and behavioural characteristic including attributive character and behavioural characteristic
To the feature of the fisrt feature shortage of data data, for example, being lacked if the fisrt feature data only include attributive character
The feature of mistake is exactly behavioural characteristic；It fills submodule 1002 and obtains historical use data to missing number in the fisrt feature data
According to feature be filled, obtain the second feature data of the active user, such as, it is assumed that active user's generic
For A class, then it is assumed that there are certain similitudes for the behavior of historical user in the behavior and A class of the active user, therefore, take A
The behavioural characteristic of historical user carries out feature filling for the fisrt feature data in class, obtains the opposite of the active user
Complete second feature data.
In an optional implementation of the present embodiment, as shown in figure 11, the filling submodule 1002 includes:
Computational submodule 1101 is configured as calculating the missing data for the historical user that active user's classification is included
Mean value；
Submodule 1102 is added, is configured as the missing data mean value that will be obtained and is added in the fisrt feature data, obtain
To the second feature data of the active user.
In this embodiment, when carrying out feature filling for the fisrt feature data using the missing data,
The mean value of the missing data of the included historical user of active user's classification can be added as feature supplementary data to described
In fisrt feature data, the relatively complete second feature data of the active user are obtained.
For example, if determining the missing data for behaviors such as visitor's unit price, preferential susceptibility, chain degree of liking through above-mentioned analysis
Operation data, and classification belonging to the active user is A class, then the visitor of all historical users in A class can be obtained
Monovalent, preferential susceptibility, it is chain like degree evidence, and it is the monovalent mean value of the visitor for calculating separately all historical users in A class, preferential quick
Sensitivity mean value and chain degree of liking mean value, are supplemented in the fisrt feature data of the active user, are worked as described
The behaviors operation datas such as visitor's unit price of preceding user, preferential susceptibility, chain degree of liking, thus form one and both belong to comprising user
Property feature, and include the second feature data of behavior operating characteristics.
Certainly, the mode of other features filling can also be used, the present invention does not make the specific implementation that feature is filled
Specific to limit, those skilled in the art can according to the needs of practical application and the characteristics of characteristic selects.
It further include for described before first prediction module 801 in an optional implementation of the present embodiment
The part that class of subscriber prediction model is trained, i.e., as shown in figure 12, described device includes:
Training module 1201, be configured as obtaining the fisrt feature data of the historical user in default historical time section and its
Classification information, and the class of subscriber is obtained according to the fisrt feature data of the historical user and the training of corresponding classification information
Prediction model；
First prediction module 1202 is configured as obtaining the fisrt feature data of active user；According to the fisrt feature
Data and class of subscriber prediction model predict the classification of the active user；
Module 1203 is filled, the characteristic pair for the historical user for being included according to active user's classification is configured as
Feature filling is carried out in the fisrt feature data, obtains the second feature data of the active user；
Second prediction module 1204, be configured as according to the second feature data for active user target data into
Row prediction.
In this embodiment, when being trained for the class of subscriber prediction model, training module 1201 is first
The fisrt feature data and its classification information for obtaining the historical user in default historical time section, then according to the historical user
Fisrt feature data and corresponding classification information training obtain the class of subscriber prediction model.
In an optional implementation of the present embodiment, the class of subscriber prediction model may be selected to be maximum entropy prediction
Model, the maximum entropy prediction model principle is to keep the entropy of prediction result maximum, i.e., so that the prediction model is estimated with unbiased
Meter.It is of course also possible to select other effective prediction models, those skilled in the art can determine suitable according to the needs of practical application
Prediction model, the present invention are not especially limited it.
In an optional implementation of the present embodiment, as shown in figure 13, second prediction module 803 includes:
Second determines submodule 1301, is configured to determine that target data prediction model；
It predicts submodule 1302, is configured as utilizing the target data prediction model according to the second feature data
The target data of active user is predicted.
In this embodiment, the target data of active user is predicted according to the second feature data
When, second determines that submodule 1301 is pre- according to the target data needed determining the characteristics of the second feature data and target data
Survey model；Submodule 1302 is predicted further according to the second feature data, using the target data prediction model for current
The target data of user is predicted.
In an optional implementation of the present embodiment, when predicting that submodule 1302 is predicted, LR mould may be selected
The prediction models such as type predict the target data, it is of course also possible to select other prediction models.Those skilled in the art
Member can select suitable prediction model according to the needs of practical application, and the present invention is not especially limited it.
In an optional implementation of the present embodiment, described device further includes being held according to the target data that prediction obtains
The part of row predetermined registration operation, i.e., as shown in figure 14, described device includes:
First prediction module 1401 is configured as obtaining the fisrt feature data of active user；According to the fisrt feature
Data and class of subscriber prediction model predict the classification of the active user；
Module 1402 is filled, the characteristic pair for the historical user for being included according to active user's classification is configured as
Feature filling is carried out in the fisrt feature data, obtains the second feature data of the active user；
Second prediction module 1403, be configured as according to the second feature data for active user target data into
Row prediction；
Execution module 1404 is configured as executing predetermined registration operation according to the target data that prediction obtains.
After prediction obtains the target data, predetermined registration operation can be executed according to the target data, wherein described default
Operation includes one of following operation or a variety of: by the target data in the current of current interface or current interface user
Operating area shows, sends or be pushed to described active user etc. for the target data.
The disclosure also discloses a kind of electronic equipment, and Figure 15 shows the knot of the electronic equipment according to one embodiment of the disclosure
Structure block diagram, as shown in figure 15, the electronic equipment 1500 include memory 1501 and processor 1502；Wherein,
The memory 1501 is for storing one or more computer instruction, wherein one or more computer
Instruction is executed by the processor 1502 to realize any of the above-described method and step.
Figure 16 is suitable for being used to realize the knot of the computer system of the target data prediction technique according to disclosure embodiment
Structure schematic diagram.
As shown in figure 16, computer system 1600 include central processing unit (CPU) 1601, can according to be stored in only
It reads the program in memory (ROM) 1602 or is loaded into random access storage device (RAM) 1603 from storage section 1608
Program and execute the various processing in above embodiment.In RAM1603, be also stored with system 1600 operate it is required various
Program and data.CPU1601, ROM1602 and RAM1603 are connected with each other by bus 1604.Input/output (I/O) interface
1605 are also connected to bus 1604.
I/O interface 1605 is connected to lower component: the importation 1606 including keyboard, mouse etc.；Including such as cathode
The output par, c 1607 of ray tube (CRT), liquid crystal display (LCD) etc. and loudspeaker etc.；Storage section including hard disk etc.
1608；And the communications portion 1609 of the network interface card including LAN card, modem etc..Communications portion 1609 passes through
Communication process is executed by the network of such as internet.Driver 1610 is also connected to I/O interface 1605 as needed.It is detachable to be situated between
Matter 1611, such as disk, CD, magneto-optic disk, semiconductor memory etc. are mounted on as needed on driver 1610, so as to
In being mounted into storage section 1608 as needed from the computer program read thereon.
Particularly, according to embodiment of the present disclosure, method as described above may be implemented as computer software programs.
For example, embodiment of the present disclosure includes a kind of computer program product comprising be tangibly embodied in and its readable medium on
Computer program, the computer program includes program code for executing above-mentioned target data prediction technique.In this way
Embodiment in, which can be downloaded and installed from network by communications portion 1609, and/or from removable
Medium 1611 is unloaded to be mounted.
Flow chart and block diagram in attached drawing illustrate system, method and computer according to the various embodiments of the disclosure
The architecture, function and operation in the cards of program product.In this regard, each box in course diagram or block diagram can be with
A part of a module, section or code is represented, a part of the module, section or code includes one or more
Executable instruction for implementing the specified logical function.It should also be noted that in some implementations as replacements, institute in box
The function of mark can also occur in a different order than that indicated in the drawings.For example, two boxes succeedingly indicated are practical
On can be basically executed in parallel, they can also be executed in the opposite order sometimes, and this depends on the function involved.Also it wants
It is noted that the combination of each box in block diagram and or flow chart and the box in block diagram and or flow chart, Ke Yiyong
The dedicated hardware based system of defined functions or operations is executed to realize, or can be referred to specialized hardware and computer
The combination of order is realized.
Being described in unit or module involved in disclosure embodiment can be realized by way of software, can also
It is realized in a manner of through hardware.Described unit or module also can be set in the processor, these units or module
Title do not constitute the restriction to the unit or module itself under certain conditions.
As on the other hand, the disclosure additionally provides a kind of computer readable storage medium, the computer-readable storage medium
Matter can be computer readable storage medium included in device described in above embodiment；It is also possible to individualism,
Without the computer readable storage medium in supplying equipment.Computer-readable recording medium storage has one or more than one journey
Sequence, described program is used to execute by one or more than one processor is described in disclosed method.
Above description is only the preferred embodiment of the disclosure and the explanation to institute's application technology principle.Those skilled in the art
Member is it should be appreciated that invention scope involved in the disclosure, however it is not limited to technology made of the specific combination of above-mentioned technical characteristic
Scheme, while should also cover in the case where not departing from the inventive concept, it is carried out by above-mentioned technical characteristic or its equivalent feature
Any combination and the other technical solutions formed.Such as features described above has similar function with (but being not limited to) disclosed in the disclosure
Can technical characteristic replaced mutually and the technical solution that is formed.
The present disclosure discloses A1, a kind of target data prediction technique, comprising: obtains the fisrt feature data of active user；
The classification of the active user is predicted according to the fisrt feature data and class of subscriber prediction model；According to the active user
The characteristic for the historical user that classification is included carries out feature filling for the fisrt feature data, obtains the current use
The second feature data at family；The target data of active user is predicted according to the second feature data.A2, according to A1
The method, the characteristic of the historical user for being included according to active user's classification is for the fisrt feature
Data carry out feature filling, obtain the second feature data of the active user, comprising: determine in the fisrt feature data and lack
Lose the feature of data；Historical use data is obtained to be filled the feature of missing data in the fisrt feature data.A3, root
According to method described in A2, the acquisition historical use data fills out the feature of missing data in the fisrt feature data
It fills, comprising: calculate the mean value of the missing data for the historical user that active user's classification is included；The missing data that will be obtained
Mean value is added in the fisrt feature data, obtains the second feature data of the active user.A4, according to any institute of A1-A3
The method stated, before the fisrt feature data for obtaining active user, further includes: obtain the history in default historical time section
The fisrt feature data and its classification information of user, and believed according to the fisrt feature data of the historical user and corresponding classification
Breath training obtains the class of subscriber prediction model.A5, according to any method of A1-A4, the class of subscriber predicts mould
Type is maximum entropy model.A6, according to any method of A1-A5, it is described according to the second feature data for currently using
The target data at family is predicted, comprising: determines target data prediction model；According to the second feature data, using described
Target data prediction model predicts the target data of active user.A7, the method according to A6, the number of targets
It is predicted that model is LR model.A8, the method according to A1-A7, further includes: executed according to the target data that prediction obtains pre-
If operation.
The present disclosure discloses B9, a kind of target data prediction meanss, comprising: the first prediction module is configured as obtaining and work as
The fisrt feature data of preceding user；Predict the active user's according to the fisrt feature data and class of subscriber prediction model
Classification；Module is filled, is configured as the characteristic for the historical user for being included according to active user's classification for described
Fisrt feature data carry out feature filling, obtain the second feature data of the active user；Second prediction module, is configured as
The target data of active user is predicted according to the second feature data.B10, the device according to B9, it is described
Filling module includes: the first determining submodule, is configured to determine that the feature of missing data in the fisrt feature data；Filling
Submodule is configured as acquisition historical use data and is filled to the feature of missing data in the fisrt feature data.
B11, device according to b10, the filling submodule includes: computational submodule, is configured as calculating the active user
The mean value of the missing data for the historical user that classification is included；Submodule is added, is configured as the missing data mean value that will be obtained
It is added in the fisrt feature data, obtains the second feature data of the active user.It is B12, any described according to B9-B11
Device, further includes: training module, be configured as obtaining the fisrt feature data of the historical user in default historical time section and
Its classification information, and the user class is obtained according to the fisrt feature data of the historical user and the training of corresponding classification information
Other prediction model.B13, according to any device of B9-B12, the class of subscriber prediction model is maximum entropy model.B14,
According to any device of B9-B13, second prediction module includes: the second determining submodule, is configured to determine that target
Data prediction model；It predicts submodule, is configured as utilizing the target data prediction model according to the second feature data
The target data of active user is predicted.B15, device according to b14, the target data prediction model are LR
Model.B16, the device according to B9-B15, further includes: execution module is configured as the target data obtained according to prediction
Execute predetermined registration operation.
The present disclosure discloses C17, a kind of electronic equipment, including memory and processor；Wherein, the memory is for depositing
Store up one or more computer instruction, wherein one or more computer instruction is executed by the processor to realize A1-
The described in any item method and steps of A8.
The disclosure also discloses D18, a kind of computer readable storage medium, is stored thereon with computer instruction, the calculating
Machine instruction realizes the described in any item method and steps of A1-A8 when being executed by processor.
Claims (10)
1. a kind of target data prediction technique characterized by comprising
Obtain the fisrt feature data of active user；According to the fisrt feature data and the prediction of class of subscriber prediction model
The classification of active user；
The characteristic for the historical user for being included according to active user's classification carries out the fisrt feature data special
Sign filling, obtains the second feature data of the active user；
The target data of active user is predicted according to the second feature data.
2. the method according to claim 1, wherein the history for being included according to active user's classification
The characteristic of user carries out feature filling for the fisrt feature data, obtains the second feature number of the active user
According to, comprising:
Determine the feature of missing data in the fisrt feature data；
Historical use data is obtained to be filled the feature of missing data in the fisrt feature data.
3. according to the method described in claim 2, it is characterized in that, the acquisition historical use data is to the fisrt feature number
It is filled according to the feature of middle missing data, comprising:
Calculate the mean value of the missing data for the historical user that active user's classification is included；
Obtained missing data mean value is added in the fisrt feature data, the second feature number of the active user is obtained
According to.
4. method according to claim 1 to 3, which is characterized in that the fisrt feature data for obtaining active user
Before, further includes:
The fisrt feature data and its classification information of the historical user in default historical time section are obtained, and are used according to the history
The fisrt feature data at family and the training of corresponding classification information obtain the class of subscriber prediction model.
5. a kind of target data prediction meanss characterized by comprising
First prediction module is configured as obtaining the fisrt feature data of active user；According to the fisrt feature data and use
The classification of active user described in the class prediction model prediction of family；
Module is filled, is configured as the characteristic of the historical user for being included according to active user's classification for described the
One characteristic carries out feature filling, obtains the second feature data of the active user；
Second prediction module is configured as predicting the target data of active user according to the second feature data.
6. device according to claim 5, which is characterized in that the filling module includes:
First determines submodule, is configured to determine that the feature of missing data in the fisrt feature data；
Submodule is filled, is configured as obtaining feature progress of the historical use data to missing data in the fisrt feature data
Filling.
7. device according to claim 6, which is characterized in that the filling submodule includes:
Computational submodule is configured as calculating the mean value of the missing data for the historical user that active user's classification is included；
Submodule is added, is configured as the missing data mean value that will be obtained and is added in the fisrt feature data, obtains described work as
The second feature data of preceding user.
8. according to any device of claim 5-7, which is characterized in that further include:
Training module is configured as obtaining the fisrt feature data of the historical user in default historical time section and its classification letter
Breath, and the class of subscriber is obtained according to the fisrt feature data of the historical user and the training of corresponding classification information and predicts mould
Type.
9. a kind of electronic equipment, which is characterized in that including memory and processor；Wherein,
The memory is for storing one or more computer instruction, wherein one or more computer instruction is by institute
Processor is stated to execute to realize the described in any item method and steps of claim 1-4.
10. a kind of computer readable storage medium, is stored thereon with computer instruction, which is characterized in that the computer instruction quilt
Claim 1-4 described in any item method and steps are realized when processor executes.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811583663.5A CN109684549A (en) | 2018-12-24 | 2018-12-24 | Target data prediction method and device, electronic equipment and computer storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811583663.5A CN109684549A (en) | 2018-12-24 | 2018-12-24 | Target data prediction method and device, electronic equipment and computer storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109684549A true CN109684549A (en) | 2019-04-26 |
Family
ID=66188204
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811583663.5A Pending CN109684549A (en) | 2018-12-24 | 2018-12-24 | Target data prediction method and device, electronic equipment and computer storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109684549A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112527852A (en) * | 2021-02-07 | 2021-03-19 | 北京淇瑀信息科技有限公司 | User dynamic support strategy allocation method and device and electronic equipment |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104598624A (en) * | 2015-02-04 | 2015-05-06 | 苏州大学 | User class determination method and device for microblog user |
| US20170329856A1 (en) * | 2015-04-08 | 2017-11-16 | Tencent Technology (Shenzhen) Company Limited | Method and device for selecting data content to be pushed to terminal, and non-transitory computer storage medium |
| CN107578294A (en) * | 2017-09-28 | 2018-01-12 | 北京小度信息科技有限公司 | User's behavior prediction method, apparatus and electronic equipment |
| CN108009926A (en) * | 2017-12-07 | 2018-05-08 | 上海点融信息科技有限责任公司 | Method, information processor and readable storage medium storing program for executing for user's classification |
| CN108830460A (en) * | 2018-05-23 | 2018-11-16 | 重庆邮电大学 | A method of recommender system Deta sparseness is alleviated based on substep dynamic filling |
| CN109033103A (en) * | 2017-06-09 | 2018-12-18 | 华为技术有限公司 | content recommendation method and system |
-
2018
- 2018-12-24 CN CN201811583663.5A patent/CN109684549A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104598624A (en) * | 2015-02-04 | 2015-05-06 | 苏州大学 | User class determination method and device for microblog user |
| US20170329856A1 (en) * | 2015-04-08 | 2017-11-16 | Tencent Technology (Shenzhen) Company Limited | Method and device for selecting data content to be pushed to terminal, and non-transitory computer storage medium |
| CN109033103A (en) * | 2017-06-09 | 2018-12-18 | 华为技术有限公司 | content recommendation method and system |
| CN107578294A (en) * | 2017-09-28 | 2018-01-12 | 北京小度信息科技有限公司 | User's behavior prediction method, apparatus and electronic equipment |
| CN108009926A (en) * | 2017-12-07 | 2018-05-08 | 上海点融信息科技有限责任公司 | Method, information processor and readable storage medium storing program for executing for user's classification |
| CN108830460A (en) * | 2018-05-23 | 2018-11-16 | 重庆邮电大学 | A method of recommender system Deta sparseness is alleviated based on substep dynamic filling |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112527852A (en) * | 2021-02-07 | 2021-03-19 | 北京淇瑀信息科技有限公司 | User dynamic support strategy allocation method and device and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105701216B (en) | A kind of information-pushing method and device | |
| CN111444966A (en) | Media information classification method and device | |
| US20200286154A1 (en) | Utilizing item-level importance sampling models for digital content selection policies | |
| US9122995B2 (en) | Classification of stream-based data using machine learning | |
| US20130138663A1 (en) | System or Apparatus for Finding Influential Users | |
| CN103412918A (en) | Quality of service (QoS) and reputation based method for evaluating service trust levels | |
| CN109033408A (en) | Information-pushing method and device, computer readable storage medium, electronic equipment | |
| Farshidi et al. | A decision support system for cloud service provider selection problem in software producing organizations | |
| CN109558544B (en) | Sorting method and device, server and storage medium | |
| CN109903095A (en) | Data processing method, device, electronic equipment and computer readable storage medium | |
| CN109697637A (en) | Object type determination method and device, electronic equipment and computer storage medium | |
| CN112418341A (en) | Model fusion method, prediction method, device, equipment and storage medium | |
| US7613799B2 (en) | Service evaluation method, system, and computer program product | |
| CN109978575A (en) | A kind of method and device excavated customer flow and manage scene | |
| JP2008158748A (en) | Variable selection device and method, and program | |
| CN111552835A (en) | File recommendation method and device and server | |
| CN108171570A (en) | A kind of data screening method, apparatus and terminal | |
| CN104077288B (en) | Web page contents recommend method and web page contents recommendation apparatus | |
| CN114297478A (en) | Page recommendation method, device, equipment and storage medium | |
| CN111209105A (en) | Capacity expansion processing method, capacity expansion processing device, capacity expansion processing equipment and readable storage medium | |
| CN117746546A (en) | Service business handling method and system based on number calling device | |
| CN109684549A (en) | Target data prediction method and device, electronic equipment and computer storage medium | |
| US11275756B2 (en) | System for extracting, categorizing and analyzing data for training user selection of products and services, and a method thereof | |
| US20140316959A1 (en) | Estimating financial risk based on non-financial data | |
| KR101981616B1 (en) | Method of providing services that predict the performance of influencer marketing |
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: 20190426 |
|
| RJ01 | Rejection of invention patent application after publication |