CN108345419A - A kind of generation method and device of information recommendation list - Google Patents
A kind of generation method and device of information recommendation list Download PDFInfo
- Publication number
- CN108345419A CN108345419A CN201710056067.0A CN201710056067A CN108345419A CN 108345419 A CN108345419 A CN 108345419A CN 201710056067 A CN201710056067 A CN 201710056067A CN 108345419 A CN108345419 A CN 108345419A
- Authority
- CN
- China
- Prior art keywords
- recommended
- characteristic
- recommendation
- target
- default
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
- G06F3/048—Interaction techniques based on graphical user interfaces [GUI]
- G06F3/0481—Interaction techniques based on graphical user interfaces [GUI] based on specific properties of the displayed interaction object or a metaphor-based environment, e.g. interaction with desktop elements like windows or icons, or assisted by a cursor's changing behaviour or appearance
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Human Computer Interaction (AREA)
- Data Mining & Analysis (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Embodiments herein discloses a kind of generation method and device of information recommendation list, the method includes:Terminal obtains the characteristic of each object to be recommended in object set to be recommended, and the object set to be recommended includes S1 objects to be recommended;Terminal, which obtains, presets characteristic set, and the default characteristic set includes the characteristic of S2 recommended, and the characteristic of S2 recommended includes the characteristic of specified recommended, and S2 is less than or equal to S1;Terminal calculates the recommendation of each object to be recommended according to the characteristic of the default characteristic set and S1 objects to be recommended;Terminal chooses target recommended according to the recommendation of each object to be recommended in S1 objects to be recommended, and target recommended is added to the appointment display position in recommendation list.Using embodiments herein, the advantages of there is the selection accuracy that recommended can be improved, improve the resource utilization of recommendation list.
Description
Technical field
This application involves field of communication technology more particularly to a kind of generation methods and device of information recommendation list.
Background technology
The relationship of the daily life demand of currently becoming increasingly popular with mobile phone, mobile phone and mobile phone user is increasingly close.Hand
Machine user can be contacted by mobile phone and kith and kin, be read, check information, planning stroke of going out, or object for appreciation game etc., and mobile phone user can
The modes such as corresponding application program (application, abbreviation APP), which are downloaded, by mobile phone application market obtains corresponding information.
Mobile phone commending system can recommend APP etc. to be downloaded on the information recommendations platform such as mobile phone application market for mobile phone user.
In the prior art, mobile phone commending system can recommend more correlations according to the APP that mobile phone user has downloaded to user
The information such as APP.The prior art is by predicting that mobile phone user to the fancy grade of certain articles, will predict that fancy grade is higher
APP recommends mobile phone user.However, the way of recommendation of the prior art only considers the information such as the interested APP of user, not
Consider the correlations such as the substitutability between the information such as the information such as existing APP and the APP of recommendation, it is poor for applicability.If for example, hand
Machine user has downloaded the APP of a GT grand touring, and the higher APP of user preferences degree predicted with this will be included more by the prior art
The APP of more GT grand tourings, therefore by the replaceable APP of more the same category in recommendation list.The GT grand touring APP that user needs is only
When needing one, the APP of other relevant GT grand tourings shows position by more recommendation is occupied, and wastes recommendation list resource,
It is poor for applicability.
Invention content
The application provides a kind of generation method and device of information recommendation list, can be improved the recommendation of object to be recommended with
The correlation of the characteristic of object to be recommended improves selection accuracy and the utilization of resources of recommendation list of recommended
Rate.
In a first aspect, this application provides a kind of generation method of information recommendation list, may include:
Terminal obtains the characteristic of each object to be recommended in object set to be recommended, in the object set to be recommended
Including S1 objects to be recommended, S1 is the integer more than 1;
The terminal, which obtains, presets characteristic set, and the default characteristic set includes S2 recommended
Characteristic, the characteristic of the S2 recommended include the characteristic of specified recommended, and S2 is less than or waits
In S1;
The terminal calculates each according to the characteristic of the default characteristic set and the S1 objects to be recommended
The recommendation of a object to be recommended;
The terminal chooses target recommendation pair according to the recommendation of each object to be recommended in the S1 objects to be recommended
As, and the target recommended is added to the appointment display position in recommendation list.
Characteristic that can be according to each object to be recommended in realization method provided by the present application and default feature
Data acquisition system calculates the recommendation of each object to be recommended.It can be by specified recommendation pair when calculating the recommendation of each object to be recommended
The characteristic of elephant also will be added in the calculating of the recommendation of object to be recommended to be determined, specify the determination of recommended can
It is determined according to practical application scene, operation is more flexible, and the calculating accuracy higher of the recommendation of object to be recommended improves generation
The algorithm complexity controllability of recommendation list, applicability are stronger.Further, realization method described herein is according to specified
The characteristic of recommended determines object to be recommended to be determined, can be between each object to be recommended of more flexible control
Correlation improves the displaying position utilization rate of recommendation list.
With reference to first aspect, in the first possible implementation, the specified recommended includes having determined to add
Add to the recommended in the recommendation list, and the characteristic similarity of the specified recommended and the target recommended
Less than default similarity threshold.
The application can flexibly select specified recommended, and the characteristic of specified recommended is added to be determined wait for
In the calculating of the recommendation of recommended, the correlation being added between each object to be recommended of recommendation list can be flexibly controlled
Property.Wherein, specified recommended may include having determined the recommended for being added to recommendation list, can reduce and alternatively wait pushing away
The probability of occurrence for recommending object further increases the effective rate of utilization of the displaying position of recommendation list, improves resource utilization.
With reference to first aspect or first aspect the first possible realization method, in second of possible realization method,
The terminal is calculated according to the characteristic of the default characteristic set and the S1 objects to be recommended described in i-th
The recommendation of object to be recommended includes:
The terminal is using the default characteristic set and the characteristic of i-th of object to be recommended as preset
Recommendation computation model input value, the recommendation of described i-th object to be recommended is calculated by the recommendation computation model
Value.
The application can calculate the recommendation of each object to be recommended by recommendation computation model, and object to be recommended can be improved
Recommendation calculating accuracy and computational efficiency, applicability higher.
With reference to first aspect to any in second of possible realization method of first aspect, in the third possible realization
In mode, the terminal chooses target recommended according to the recommendation of each object to be recommended in the S1 objects to be recommended
Including:
The terminal chooses the object to be recommended for meeting predefined selection rule from the S1 objects to be recommended, and from
It is target recommended that the maximum object to be recommended of recommendation is chosen in the object to be recommended for meeting predefined selection rule.
The application can choose target recommended by predefining the recommendation of selection rule and each object to be recommended, into
One step improves the selection controllability of each recommended in recommendation list, improves the resource utilization of recommendation list.
It is any in the third possible realization method to first aspect with reference to first aspect, in the 4th kind of possible realization
In mode, it is described the target recommended is added to the appointment display position in recommendation list after, the method further includes:
The characteristic of the target recommended is rejected from the object set to be recommended, and the target is pushed away
The characteristic for recommending object is added in the default characteristic set.
The application can preset the characteristic in characteristic set by the target recommended update that basis has been chosen,
It improves the selection controllability of each recommended in recommendation list, reduces in recommendation list going out for interchangeable object to be recommended
Existing probability, improves the resource utilization of recommendation list.
With reference to first aspect to any in the 4th kind of possible realization method of first aspect, in the 5th kind of possible realization
In mode, the object to be recommended includes:At least one in application APP, audio, video data, webpage and Domestic News
Kind.
Realization method provided by the present application is applicable to the selection of the object to be recommended of more forms of expression, improves information
The diversity of recommendation list, the applicability of the generation method of enhancement information recommendation list.
With reference to first aspect to any in the 5th kind of possible realization method of first aspect, in the 6th kind of possible realization
In mode, the characteristic includes:Identity ID, category attribute are applicable in platform, consult number, click-through-rate, download
At least one of number and size of data.
In realization method provided by the present application, the spy in the characteristic of object to be recommended and default characteristic set
Sign data may include a plurality of types of data, and the selection accuracy of object to be recommended can be improved, enhance the validity of recommendation list.
It is any in the third possible realization method to first aspect with reference to first aspect, in the 7th kind of possible realization
In mode, the predefined rule includes:The number of the same or similar object to be recommended of feature is not more than M1, or waits pushing away
The version updating date for recommending object is not later than the predefined date.
The application can choose target recommended by predefining the recommendation of selection rule and each object to be recommended, into
One step improves the selection validity of each recommended in recommendation list, improves the resource utilization of recommendation list.
Second aspect, this application provides a kind of generating means of information recommendation list, may include:
Acquisition module, the characteristic for obtaining each object to be recommended in object set to be recommended are described to be recommended
Object set includes S1 objects to be recommended, and S1 is the integer more than 1;
The acquisition module is additionally operable to obtain default characteristic set, and the default characteristic set includes S2
The characteristic of a recommended, the characteristic of the S2 recommended include the characteristic of specified recommended,
S2 is less than or equal to S1;
Computing module, the default characteristic set and the S1 for being obtained according to the acquisition module wait pushing away
The characteristic of object is recommended, the recommendation of each object to be recommended is calculated;
Module is chosen, it is each to be recommended in the S1 for being calculated according to the computing module object to be recommended
The recommendation of object chooses target recommended, and the target recommended is added to the appointment display in recommendation list
Position.
In conjunction with second aspect, in the first possible implementation, the specified recommendation that the acquisition module obtains
Object includes having determined the recommended being added in the recommendation list, and the specified recommended is pushed away with the target
The characteristic similarity for recommending object is less than default similarity threshold.
In conjunction with second aspect or second aspect the first possible realization method, in second of possible realization method,
The computing module is used for:
Using the default characteristic set and the characteristic of i-th of object to be recommended as preset recommendation
The input value of computation model calculates the recommendation of described i-th object to be recommended by the recommendation computation model.
In conjunction with any in second aspect to second of possible realization method of second aspect, in the third possible realization
In mode, the selection module is used for:
It chooses from the S1 objects to be recommended that the acquisition module obtains and to meet predefined selection rule and wait pushing away
Object is recommended, and from the object to be recommended for meeting predefined selection rule, that chooses that the computing module is calculated pushes away
It is target recommended to recommend the maximum object to be recommended of value.
In conjunction with any in second aspect to second aspect the third possible realization method, in the 4th kind of possible realization
In mode, the acquisition module is additionally operable to:
The characteristic of the target recommended is rejected from the object set to be recommended, and the target is pushed away
The characteristic for recommending object is added in the default characteristic set.
In conjunction with any in the 4th kind of possible realization method of second aspect to second aspect, in the 5th kind of possible realization
In mode, the object to be recommended includes:At least one in application APP, audio, video data, webpage and Domestic News
Kind.
In conjunction with any in the 5th kind of possible realization method of second aspect to second aspect, in the 6th kind of possible realization
In mode, the characteristic includes:Identity ID, category attribute are applicable in platform, consult number, click-through-rate, download
At least one of number and size of data.
The third aspect, this application provides a kind of terminal devices, may include:Memory and processor;
The memory is for storing batch processing code;
The processor is used to that the program code stored in the memory to be called to execute the side that above-mentioned first aspect provides
Method.
Characteristic that can be according to each object to be recommended in realization method provided by the present application and default feature
Data acquisition system calculates the recommendation of each object to be recommended, wherein has determined the characteristic of the object to be recommended of recommendation
It will be added in the calculating of the recommendation of next object to be recommended to be determined, the calculating accuracy higher of recommendation, improve
The algorithm complexity controllability of recommendation list is generated, applicability is stronger.Further, realization method root described herein
Object to be recommended to be determined is determined according to the characteristic for the object to be recommended having determined, more flexible can control each wait for
Correlation between recommended reduces the probability of occurrence of alternatively object to be recommended, improves the displaying position profit of recommendation list
With rate.
Description of the drawings
It, below will be to institute in embodiments herein in order to illustrate more clearly of the technical solution in embodiments herein
Attached drawing to be used is needed to illustrate.
Fig. 1 is a flow diagram of the generation method for the information recommendation list that embodiments herein provides;
Fig. 2 is another flow diagram of the generation method for the information recommendation list that embodiments herein provides;
Fig. 3 is a structural schematic diagram of the generating means for the information recommendation list that embodiments herein provides;
Fig. 4 is another structural schematic diagram of the generating means for the information recommendation list that embodiments herein provides.
Specific implementation mode
Embodiments herein is described with reference to the attached drawing in embodiments herein.
In the specific implementation, the terminal described in embodiments herein includes:Mobile phone (mobile phone), tablet electricity
Brain (Pad), wearable device and personal computer assistant etc., can specifically determine according to practical application scene, not limit herein
System.It will be illustrated by taking mobile phone as an example below.
The generation method and the application platform that is applicable in of device for the information recommendation list that embodiments herein provides include
But be not limited to mobile phone application market, mobile phone music, mobile video, mobile phone reading, mobile phone Domestic News and mobile phone web pages etc..Example
Such as, by taking application market as an example, the generation method and device of the information recommendation list of embodiments herein offer are applicable to China
For application market, Baidu's application market, 360 application markets and millet application market etc..Described in embodiments herein
Information recommendation list may include but be not limited to:APP recommendation lists, audio, video data recommendation list, webpage recommending list and new
Hear information recommendation list etc..That is, object to be recommended described herein may include but be not limited to:APP, audio, video data, net
Page and Domestic News etc., will illustrate by taking APP as an example below.
In mobile phone commending system, the generation of information recommendation list includes the processes such as prediction and recommendation.Wherein, needed for prediction
What is solved is the fancy grade for predicting mobile phone user to each recommended.Recommendation is then will recommendation pair according to the result of prediction
As being ranked up, such as according to the fancy grade of prediction, it is ranked up according to the high to Low sequence of fancy grade.Sequence study
The ordering strategy that the field (learning to rank) proposes includes point-wise (single spot optimization) strategies and list-wise
(list optimization) strategy etc..Wherein, above-mentioned point-wise strategies be according in object to be recommended each object to be recommended it is pre-
Recommendation is surveyed, such as click-through-rate (click-through-rate, CTR), sequence from big to small are ranked up.
In point-wise strategies, the sequence of each object to be recommended is according to the user preferences degree of each object to be recommended, from height
It is ranked up to low fancy grade, does not consider the correlations such as the substitutability between each object to be recommended, it is poor for applicability.
List-wise ordering strategies are then directly to be ranked up all recommendeds to obtain a whole sequence, in turn
Using the Unitary serial as a sample, directly optimize the Unitary serial and obtain the sequence after an optimization, by all recommendations pair
As being ranked up to obtain recommendation list according to the sequence after optimization.The difficult point of the ordering strategy is how to be carried out to recommendation list
Mark, and need to calculate the probability of all sequence combinations, realize that difficulty is big.If recommended has n, the ordering strategy institute
The time complexity needed will be up to O (n*n!), wherein O () is the expression formula of time complexity.N is bigger, then the collating sequence
Required time complexity is higher, can not directly apply to and solve the problems, such as industrial quarters, poor for applicability.
Embodiments herein provides a kind of generation method and device of information recommendation list, can be according to each to be recommended right
The feature of the feature of elephant and ordering recommended chooses target recommended and is added to the finger in recommendation list successively
In fixed displaying position.The realization method that embodiments herein provides is contemplated that the correlations such as the replaceability between each recommended
Property, the resource profit of recommendation list is improved in the displaying position for the occupancy recommendation list for avoiding the higher recommended of replaceability excessive
With rate, the applicability of the generation of information recommendation list is improved.
It is a flow diagram of the generation method for the information recommendation list that embodiments herein provides referring to Fig. 1.This
The method that the embodiment of application provides includes:
S101, mobile phone obtain the characteristic of each object to be recommended in object set to be recommended.
In some feasible embodiments, mobile phone can obtain the feature of object to be recommended from platforms such as application markets first
Data, and object set F1 to be recommended is obtained according to the characteristic of all objects to be recommended acquired.Wherein, above-mentioned F1
The object number to be recommended for including may be set to S1, and S1 is the integer more than 1.Mobile phone generates APP recommendation lists (hereinafter referred to as
Recommendation list) when the recommendation of each object to be recommended can be determined according to the characteristic of above-mentioned each object to be recommended, in turn
Target recommended can be chosen according to the recommendation of each object to be recommended, and is exported to the appointment display position of recommendation list.
It, can be first before mobile phone generates APP recommendation lists (hereinafter referred to as recommendation list) in some feasible embodiments
Recommendation list is initialized, recommendation list (list) is initialized as sky.Further, mobile phone executes the first of recommendation list
When beginningization operates, the initial characteristics set F2 needed for the recommendation for calculating each object to be recommended can be set, can also be set
Ranked object set F3.Wherein, above-mentioned F2 includes the initial characteristics number of the recommendation for calculating each object to be recommended
According to above-mentioned initial characteristics data include:Characteristic, the use of each object to be recommended in S1 objects to be recommended of above-mentioned acquisition
The assemblage characteristic etc. of family characteristic and the characteristic and the user characteristic data of above-mentioned each object to be recommended, specifically
It can be determined according to practical application scene or application platform, be not limited herein.Wherein, the feature of above-mentioned each object to be recommended
Data may include but be not limited to:Identity (identity, ID), product attribute, category attribute, be applicable in platform, consult number,
Click-through-rate, download time and size of data etc. are not limited this application.Above-mentioned user characteristic data may include but
It is not limited to:User ID, user select APP and history downloading data or historical viewings data etc., do not limit this application
System.
Further, above-mentioned F3 is sky when initialization is completed, and mobile phone, which is often handled, to be obtained a recommended and sorted
To the appointment display position of recommendation list, then ordering recommended can be added in F3, and then can be wrapped according in F3
The number of the object to be recommended (i.e. ranked object) included determines the need for choosing new target recommended.If F3 includes
Ranked object number be more than or equal to predefined number, then no longer choose new target recommended, and according to
The object of sequence generates recommendation list.
S102, mobile phone, which obtains, presets characteristic set.
In some feasible embodiments, when mobile phone determines the first aim recommended of recommendation list, acquisition
Default characteristic set can be the initial characteristics set F2 that above-mentioned initialization obtains.That is, in the application scenarios, feature is preset
Characteristic included by data acquisition system is identical as the characteristic that above-mentioned initial characteristics set F2 includes.
Further, in some feasible embodiments, when mobile phone determines other target recommendeds of recommendation list,
The characteristic of the default characteristic set obtained may include the characteristic that above-mentioned F2 includes, and one had determined
The characteristic of a or multiple (being set as S2) recommended.Wherein, above-mentioned S2 is more than 1 and is less than or equal to S1.
S103, mobile phone are calculated according to the characteristic of the default characteristic set and the S1 objects to be recommended
The recommendation of each object to be recommended.
In some feasible embodiments, mobile phone calculates any object to be recommended in the corresponding objects to be recommended of above-mentioned F1
When the recommendation of (being set as object 1 to be recommended), the characteristic that can include by the default characteristic set of above-mentioned acquisition with
And the characteristic of above-mentioned object to be recommended 1, calculate the recommendation of object 1 to be recommended.In the specific implementation, mobile phone can will be to be recommended
The characteristic that the characteristic of object 1 and above-mentioned default characteristic set include calculates mould as preset recommendation
The input value of type calculates the recommendation of object 1 to be recommended by recommendation computation model.Similarly, mobile phone can be calculated above-mentioned
The recommendation for other objects to be recommended that F1 includes.
It should be noted that recommendation computation model described herein may include but be not limited to:Logistic regression
(logistic regression) model, decision-tree model, deep learning model, factorization machine model, domain perceive factor point
The built-up pattern etc. of solution machine model and above-mentioned any number of models.It can specifically be determined according to practical application scene demand, herein
It is not limited.The effect of above-mentioned recommendation computation model is to determine the recommender score (i.e. recommendation) of object to be recommended.It is above-mentioned to push away
Recommending value computation model can be trained to obtain by the generation data of history recommendation list by machine learning algorithm.Wherein, above-mentioned history
The generation data of recommendation list may include but be not limited to label (label), history recommendation list and its including recommended
Characteristic etc..Wherein, above-mentioned label is the user's operation behavior of object to be recommended, such as buys or does not buy, be clicked or
It is not clicked on.
In the specific implementation, mobile phone can be by the spy in the characteristic of above-mentioned object 1 to be recommended and default characteristic set
Sign data are combined to obtain assemblage characteristic, and then the recommendation of object 1 to be recommended can be calculated according to assemblage characteristic.Wherein,
The combination of features described above data may include feature combination of cartesian product etc..It should be noted that above-mentioned default spy
Sign data acquisition system may include the spy for the object to be recommended (i.e. specified recommended, be set as object 2 to be recommended) for having determined recommendation
Levy data.When mobile phone calculates the recommendation of object 1 to be recommended, can by the characteristic of object 2 to be recommended and object 1 to be recommended with
And the characteristic in F1 is combined, the characteristic obtained according to combination calculates the recommendation of object 1 to be recommended.
In some feasible embodiments, mobile phone calculates the feature of each object to be recommended by recommendation computation model
After data, the recommendation adjustment rule of object to be recommended can be also predefined.Wherein, above-mentioned recommendation adjustment rule can be:If
Object 1 to be recommended and the characteristic similarity of fixed recommended (i.e. ordering recommended) are more than or equal to default
Similarity threshold then lowers the recommendation of object 1 to be recommended;If the feature of object 1 and ordering recommended to be recommended
Similarity is less than default similarity threshold, then raises the recommendation of object 1 to be recommended.If for example, object to be recommended 1 with wait pushing away
The characteristic similarity for recommending object 2 is more than or equal to default similarity threshold, then can determine object 1 to be recommended and object to be recommended
2 be analogical object, and the recommendation for the object to be recommended 1 being calculated at this time can be adjusted to smaller value.If object to be recommended 1 and waiting for
The characteristic similarity of recommended 2 is less than default similarity threshold, then can determine object 1 and object to be recommended 2 to be recommended not
The recommendation of analogical object, the object to be recommended 1 being calculated at this time can be adjusted to higher value.Wherein, features described above similarity
It may include:Classification similarity, information type similarity or applicable platform similarity etc..
S104, mobile phone choose target recommendation pair according to the recommendation of each object to be recommended in the S1 objects to be recommended
As, and the target recommended is added to the appointment display position in recommendation list.
In some feasible embodiments, mobile phone is calculated in F1 after the recommendation of each object to be recommended, then
The maximum target recommended of recommendation can be chosen from the recommendation of each object to be recommended, and the target recommended is added
Add to the appointment display position in recommendation list.Wherein, above-mentioned appointment display position concretely last position in recommendation list.
That is, the target recommended that mobile phone is chosen every time is both placed in last position of recommendation list, if the target recommended chosen
Number be equal to it is predefined recommend number, then produce output to the recommendation list on user interface.
It should be noted that in realization method described herein, in the recommendation for the object to be recommended that rear 1 determines
Calculating all can be with reference to the characteristic of the ordering recommended in preceding determination, and then can enhance between each object to be recommended
Feature association, the probability for occurring replaceable recommended in recommendation list can be reduced, improve the resource utilization of recommendation list.
Further, in some feasible embodiments, mobile phone selection target recommendation pair from each object to be recommended
As when, can also according to predefined selection rule further screen output to recommendation list recommended.Wherein, above-mentioned predefined
Selection rule may include:
Rule 1:The number of the same or similar object to be recommended of feature is not more than M1.
Wherein, M1 can be defined according to practical application scene, be not limited herein.For example, it is assumed that M1 is 2, and predefine
When selection rule is that the number of the same or similar object to be recommended of classification in recommendation list is not more than 2, if ordering push away
The recommended for recommending classification A in object has had 2, then when choosing target recommended according to recommendation, even if classification is A's
The recommendation maximum of object to be recommended is not chosen to be target recommended yet.At this point, classification can be B and recommended by mobile phone
The object to be recommended that value is only smaller than maximum recommended value is determined as target recommended.And then it can avoid same or similar waiting pushing away
The displaying position resource for recommending the excessive occupancy recommendation list of object, improves the resource utilization of recommendation list.
Rule 2:The date of reaching the standard grade of object to be recommended is not later than the predefined date.
For example, if predefined selection rule regulation, the recommended of the third displaying position displaying of recommendation list can only be
The object reached the standard grade in nearest one week, then mobile phone determine third target recommended, can be from the object to be recommended not being sorted also
It is middle to choose the object to be recommended reached the standard grade in nearest one week, then the maximum target of recommendation is determined from the object to be recommended of selection
Recommended is as third target recommended.And then the object to be recommended that can avoid lowest version excessively occupies recommendation list
Displaying position resource, improve the effectiveness of information of recommendation list, enhance the applicability of recommendation list.
It further,, then can be by target after mobile phone determines target recommended in some feasible embodiments
Recommended is rejected from above-mentioned F1 and is added in F3, and then the characteristic of target recommended is added in F2, with
Fixed reference feature data as the next target recommended of determination.If the quantity of ordering recommended is more than or equal to
It is predefined to recommend number or above-mentioned F1 for sky, then recommendation list can be generated according to ordering recommended, and will recommend to arrange
Table is exported to user interface.
It is another flow chart of the generation method for the information recommendation list that embodiments herein provides referring to Fig. 2.This Shen
The cyclic process of the generation method for the information recommendation list that embodiment please provides includes:
S201, data initialization.
In the specific implementation, above-mentioned data initialization process includes above-mentioned F1, F2, F3, recommendation list, default characteristic data set
Close the initialization of the data such as the predefined selection rule for stating recommendation adjustment rule and above-mentioned target recommended.
S202 calculates the recommendation of each object to be recommended by recommendation computation model.
In the specific implementation, the calculation of the recommendation of above-mentioned each object to be recommended can be found in it is each in above-described embodiment
Realization method described in step, details are not described herein.
S203, the predefined selection rule that rule or target recommended are adjusted according to recommendation determine target recommendation pair
As.
In the specific implementation, the method for determination of above-mentioned target recommended can be found in above-described embodiment described by each step
Realization method, details are not described herein.
S204 updates F1, F2, F3 and default characteristic set.
In the specific implementation, the update mode of above-mentioned F1, F2, F3 and default characteristic set can participate in above-described embodiment
Realization method described in middle correlation step, details are not described herein.
S205, determines whether the number of object to be recommended during whether F1 is empty or F3 meets the requirements.
It is met the requirements if F1 is the number of ordering recommended in empty or F3, thens follow the steps S206, otherwise hold
Row step S202.
S206 generates recommendation list and exports to user interface.
In realization method described herein, mobile phone often determines a target recommended, then may be updated corresponding
Data acquisition system, until all objects to be recommended sort, the number of completion or ordering recommended meets predefined
Number requirement, then can ultimately generate recommendation list and export to user interface, to be presented to user.
It should be noted that if above- mentioned information recommendation list is webpage recommending list, then it is described herein to be recommended
The characteristic of object may also include but be not limited to:Web page title, webpage Main Domain, the term weighing in webpage and web page class
Not etc., it can specifically be determined according to practical application scene, be not limited herein.
In realization method described herein, it has been determined that object to be recommended characteristic will be added to it is next
In the calculating of the recommendation of a object to be recommended to be determined, the calculating accuracy higher of recommendation improves generation and recommends row
The algorithm complexity controllability of table, applicability are stronger.Further, realization method described herein is according to having determined
The characteristic of object to be recommended determines object to be recommended to be determined, can be between each object to be recommended of more flexible control
Correlation, reduce alternatively object to be recommended probability of occurrence, improve the displaying position utilization rate of recommendation list.
It is a structural schematic diagram of the generating means for the information recommendation list that embodiments herein provides referring to Fig. 3.This
The generating means for the information recommendation list that the embodiment of application the provides concretely terminal device in above-described embodiment, such as hand
Machine etc., is not limited herein.Wherein, the generating means (hereinafter referred to as terminal) of above- mentioned information recommendation list may include:
Acquisition module 31, the characteristic for obtaining each object to be recommended in object set to be recommended are described to wait pushing away
It includes S1 objects to be recommended to recommend object set, and S1 is the integer more than 1.
The acquisition module 31 is additionally operable to obtain default characteristic set, and the default characteristic set includes
The characteristic of S2 recommended, the characteristic of the S2 recommended include the characteristic of specified recommended
According to S2 is less than or equal to S1.
Computing module 32, the default characteristic set and the S1 for being obtained according to the acquisition module wait for
The characteristic of recommended calculates the recommendation of each object to be recommended.
Module 33 is chosen, it is each in the S1 for being calculated according to the computing module object to be recommended to wait pushing away
The recommendation for recommending object chooses target recommended, and the target recommended is added to the appointment display in recommendation list
Position.
In some feasible embodiments, the specified recommended that above-mentioned acquisition module 31 obtains includes really
Surely the recommended being added in the recommendation list, and the feature phase of the specified recommended and the target recommended
It is less than default similarity threshold like degree.
In some feasible embodiments, above-mentioned computing module 32 is used for:
Using the default characteristic set and the characteristic of i-th of object to be recommended as preset recommendation
The input value of computation model calculates the recommendation of described i-th object to be recommended by the recommendation computation model.
In some feasible embodiments, above-mentioned selection module 33 is used for:
It is chosen in the S1 objects to be recommended obtained from the acquisition module 31 and meets waiting for for predefined selection rule
Recommended, and from the object to be recommended for meeting predefined selection rule, choose the computing module 32 and be calculated
The maximum object to be recommended of recommendation be target recommended.
In some feasible embodiments, above-mentioned acquisition module 31 is additionally operable to:
The characteristic of the target recommended is rejected from the object set to be recommended, and the target is pushed away
The characteristic for recommending object is added in the default characteristic set.
In some feasible embodiments, above-mentioned object to be recommended includes:Application APP, audio, video data, webpage
And at least one of Domestic News.
In some feasible embodiments, features described above data include:Identity ID, category attribute, be applicable in platform,
Consult at least one of number, click-through-rate, download time and size of data.
In the specific implementation, terminal provided herein can be executed by the modules included by it in above-described embodiment
Realization method described in each step, details are not described herein.
It is another structural schematic diagram of the generating means for the information recommendation list that embodiments herein provides referring to Fig. 4.
The generating means for the information recommendation list that embodiments herein is provided can be the terminal in above-described embodiment, may include depositing
Reservoir 41 and processor 42, wherein memory 41 can be connected with processor 42 by bus.
Memory 41 include but not limited to be random access memory (random access memory, RAM), read-only deposit
Reservoir (read-only memory, ROM), Erasable Programmable Read Only Memory EPROM (erasable programmable read
Only memory, EPROM) or portable read-only memory (compact disc read-only memory, CD-ROM),
The memory 41 is used for dependent instruction and data.
Processor 42 can be one or more central processing units (central processing unit, CPU), locate
In the case that reason device 42 is a CPU, which can be monokaryon CPU, can also be multi-core CPU.
Above-mentioned processor 42 executes following operation for reading the program code stored in memory 41:
The characteristic of each object to be recommended in object set to be recommended is obtained, the object set to be recommended includes
S1 objects to be recommended, S1 are the integer more than 1;
It obtains and presets characteristic set, the default characteristic set includes the characteristic of S2 recommended
According to the characteristic of the S2 recommended includes the characteristic of specified recommended, and S2 is less than or equal to S1;
According to the characteristic of the default characteristic set and the S1 objects to be recommended, each described wait for is calculated
The recommendation of recommended;
Target recommended is chosen according to the recommendation of each object to be recommended in the S1 objects to be recommended, and by institute
State the appointment display position that target recommended is added in recommendation list.
In some feasible embodiments, above-mentioned specified recommended includes having determined to be added to the recommendation list
In recommended, and the characteristic similarity of the specified recommended and the target recommended is less than default similarity threshold
Value.
In some feasible embodiments, above-mentioned processor 42 is used for:
Using the default characteristic set and the characteristic of i-th of object to be recommended as preset recommendation
The input value of computation model calculates the recommendation of described i-th object to be recommended by the recommendation computation model.
In some feasible embodiments, above-mentioned processor 42 is used for:
The object to be recommended for meeting predefined selection rule is chosen from the S1 objects to be recommended, and is met from described
It is target recommended that the maximum object to be recommended of recommendation is chosen in the object to be recommended of predefined selection rule.
In some feasible embodiments, above-mentioned processor 42 is additionally operable to:
The characteristic of the target recommended is rejected from the object set to be recommended, and the target is pushed away
The characteristic for recommending object is added in the default characteristic set.
In some feasible embodiments, above-mentioned object to be recommended includes:Application APP, audio, video data, webpage
And at least one of Domestic News.
In some feasible embodiments, features described above data include:Identity ID, category attribute, be applicable in platform,
Consult at least one of number, click-through-rate, download time and size of data.
Characteristic that can be according to each object to be recommended in realization method provided by the present application and default feature
Data acquisition system calculates the recommendation of each object to be recommended, wherein has determined the characteristic of the object to be recommended of recommendation
It will be added in the calculating of the recommendation of next object to be recommended to be determined, the calculating accuracy higher of recommendation, improve
The algorithm complexity controllability of recommendation list is generated, applicability is stronger.Further, realization method root described herein
Object to be recommended to be determined is determined according to the characteristic for the object to be recommended having determined, more flexible can control each wait for
Correlation between recommended reduces the probability of occurrence of alternatively object to be recommended, improves the displaying position profit of recommendation list
With rate.
One of ordinary skill in the art will appreciate that realizing all or part of flow in above-described embodiment method, the flow
Relevant hardware can be instructed to complete by computer program, which can be stored in computer read/write memory medium, should
Program is when being executed, it may include such as the flow of above-mentioned each method embodiment.And storage medium above-mentioned includes:ROM is deposited at random
Store up the medium of the various program storage codes such as memory body RAM, magnetic disc or CD.
Claims (10)
1. a kind of generation method of information recommendation list, which is characterized in that including:
Terminal obtains the characteristic of each object to be recommended in object set to be recommended, and the object set to be recommended includes
S1 objects to be recommended, S1 are the integer more than 1;
The terminal, which obtains, presets characteristic set, and the default characteristic set includes the feature of S2 recommended
Data, the characteristic of the S2 recommended include the characteristic of specified recommended, and S2 is less than or equal to S1;
The terminal calculates each institute according to the characteristic of the default characteristic set and the S1 objects to be recommended
State the recommendation of object to be recommended;
The terminal chooses target recommended according to the recommendation of each object to be recommended in the S1 objects to be recommended, and
The target recommended is added to the appointment display position in recommendation list.
2. the method as described in claim 1, which is characterized in that the specified recommended include have determined be added to it is described
Recommended in recommendation list, and the characteristic similarity of the specified recommended and the target recommended is less than default
Similarity threshold.
3. method as claimed in claim 1 or 2, which is characterized in that the terminal according to the default characteristic set and
The characteristic of the S1 objects to be recommended, the recommendation for calculating i-th of object to be recommended include:
The terminal pushes away the characteristic of the default characteristic set and i-th of object to be recommended as preset
The input value for recommending value computation model calculates the recommendation of described i-th object to be recommended by the recommendation computation model.
4. method as described in any one of claims 1-3, which is characterized in that the terminal is according to the S1 objects to be recommended
In each object to be recommended recommendation choose target recommended include:
The terminal chooses the object to be recommended for meeting predefined selection rule from the S1 objects to be recommended, and from described
Meet and chooses the maximum object to be recommended of recommendation in the object to be recommended of predefined selection rule as target recommended.
5. method according to any one of claims 1-4, which is characterized in that described be added to the target recommended pushes away
After recommending the appointment display position in list, the method further includes:
The characteristic of the target recommended is rejected from the object set to be recommended, and by the target recommend pair
The characteristic of elephant is added in the default characteristic set.
6. a kind of generating means of information recommendation list, which is characterized in that including:
Acquisition module, the characteristic for obtaining each object to be recommended in object set to be recommended, the object to be recommended
Set includes S1 objects to be recommended, and S1 is the integer more than 1;
The acquisition module is additionally operable to obtain default characteristic set, and the default characteristic set includes S2 and pushes away
The characteristic of object is recommended, the characteristic of the S2 recommended includes the characteristic of specified recommended, and S2 is small
In or equal to S1;
Computing module, the default characteristic set and the S1 for being obtained according to the acquisition module are a to be recommended right
The characteristic of elephant calculates the recommendation of each object to be recommended;
Module is chosen, each object to be recommended in the S1 for being calculated according to the computing module object to be recommended
Recommendation choose target recommended, and the target recommended is added to the appointment display position in recommendation list.
7. generating means as claimed in claim 6, which is characterized in that the specified recommended that the acquisition module obtains
Including having determined the recommended being added in the recommendation list, and the specified recommended and target recommendation pair
The characteristic similarity of elephant is less than default similarity threshold.
8. generating means as claimed in claims 6 or 7, which is characterized in that the computing module is used for:
The default characteristic set and the characteristic of i-th of object to be recommended are calculated as preset recommendation
The input value of model calculates the recommendation of described i-th object to be recommended by the recommendation computation model.
9. such as claim 6-8 any one of them generating means, which is characterized in that the selection module is used for:
It is chosen in the S1 objects to be recommended obtained from the acquisition module and meets the to be recommended right of predefined selection rule
As, and from the object to be recommended for meeting predefined selection rule, choose the recommendation that the computing module is calculated
Maximum object to be recommended is target recommended.
10. such as claim 6-9 any one of them generating means, which is characterized in that the acquisition module is additionally operable to:
The characteristic of the target recommended is rejected from the object set to be recommended, and by the target recommend pair
The characteristic of elephant is added in the default characteristic set.
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710056067.0A CN108345419B (en) | 2017-01-25 | 2017-01-25 | Information recommendation list generation method and device |
PCT/CN2017/115551 WO2018137420A1 (en) | 2017-01-25 | 2017-12-12 | Generating method and device for information recommendation list |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710056067.0A CN108345419B (en) | 2017-01-25 | 2017-01-25 | Information recommendation list generation method and device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108345419A true CN108345419A (en) | 2018-07-31 |
CN108345419B CN108345419B (en) | 2021-06-08 |
Family
ID=62962843
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710056067.0A Active CN108345419B (en) | 2017-01-25 | 2017-01-25 | Information recommendation list generation method and device |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN108345419B (en) |
WO (1) | WO2018137420A1 (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109408729A (en) * | 2018-12-05 | 2019-03-01 | 广州市百果园信息技术有限公司 | Material is recommended to determine method, apparatus, storage medium and computer equipment |
CN110152298A (en) * | 2018-11-22 | 2019-08-23 | 腾讯科技(深圳)有限公司 | Realize equipment recommended method, device and the computer equipment of virtual role |
CN110827109A (en) * | 2018-08-14 | 2020-02-21 | 中国移动通信集团浙江有限公司 | Short message marketing method and device |
CN111026934A (en) * | 2019-12-20 | 2020-04-17 | 中科寒武纪科技股份有限公司 | Intelligent recommendation method and related equipment |
WO2020134011A1 (en) * | 2018-12-29 | 2020-07-02 | 北京字节跳动网络技术有限公司 | Method and apparatus for determining display information combination, storage medium, and electronic device |
CN115329208A (en) * | 2022-10-18 | 2022-11-11 | 平安银行股份有限公司 | Interest fee exemption scheme recommendation method and device, electronic equipment and medium |
Citations (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20100228653A1 (en) * | 2009-03-05 | 2010-09-09 | Fujifilm Corporation | Data reproduction device, content delivery system, and content delivery method |
CN103473244A (en) * | 2012-06-08 | 2013-12-25 | 富士通株式会社 | Device and method for recommending applications used in application group |
CN103744966A (en) * | 2014-01-07 | 2014-04-23 | Tcl集团股份有限公司 | Item recommendation method and device |
US20140310304A1 (en) * | 2013-04-12 | 2014-10-16 | Ebay Inc. | System and method for providing fashion recommendations |
CN104636419A (en) * | 2013-11-11 | 2015-05-20 | 国际商业机器公司 | Information processing device, information processing method |
US20160094501A1 (en) * | 2014-09-26 | 2016-03-31 | Line Corporation | Method, system and recording medium for providing video contents in social platform and file distribution system |
CN105469263A (en) * | 2014-09-24 | 2016-04-06 | 阿里巴巴集团控股有限公司 | Commodity recommendation method and device |
CN105812937A (en) * | 2014-12-30 | 2016-07-27 | Tcl集团股份有限公司 | Television program recommending method and television program recommending device |
CN105843953A (en) * | 2016-04-12 | 2016-08-10 | 乐视控股(北京)有限公司 | Multimedia recommendation method and device |
CN105930429A (en) * | 2016-04-19 | 2016-09-07 | 乐视控股(北京)有限公司 | Music recommendation method and apparatus |
CN105989122A (en) * | 2015-02-12 | 2016-10-05 | 广东欧珀移动通信有限公司 | Application recommendation method and device |
CN106021485A (en) * | 2016-05-19 | 2016-10-12 | 中国传媒大学 | Multi-element attribute movie data visualization system |
CN106227815A (en) * | 2015-07-22 | 2016-12-14 | Tcl集团股份有限公司 | The personalized application program function of a kind of multi-modal clue recommends method and system thereof |
CN106326431A (en) * | 2016-08-25 | 2017-01-11 | 乐视控股(北京)有限公司 | Information recommendation method and device |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP4247280B2 (en) * | 2007-02-14 | 2009-04-02 | 株式会社東芝 | RECOMMENDED PROGRAM INFORMATION PROVIDING DEVICE, RECOMMENDED PROGRAM INFORMATION PROVIDING METHOD, AND PROGRAM |
CN102831234B (en) * | 2012-08-31 | 2015-04-22 | 北京邮电大学 | Personalized news recommendation device and method based on news content and theme feature |
CN105608117B (en) * | 2015-12-14 | 2019-12-10 | 微梦创科网络科技(中国)有限公司 | Information recommendation method and device |
-
2017
- 2017-01-25 CN CN201710056067.0A patent/CN108345419B/en active Active
- 2017-12-12 WO PCT/CN2017/115551 patent/WO2018137420A1/en active Application Filing
Patent Citations (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20100228653A1 (en) * | 2009-03-05 | 2010-09-09 | Fujifilm Corporation | Data reproduction device, content delivery system, and content delivery method |
CN103473244A (en) * | 2012-06-08 | 2013-12-25 | 富士通株式会社 | Device and method for recommending applications used in application group |
US20140310304A1 (en) * | 2013-04-12 | 2014-10-16 | Ebay Inc. | System and method for providing fashion recommendations |
CN104636419A (en) * | 2013-11-11 | 2015-05-20 | 国际商业机器公司 | Information processing device, information processing method |
CN103744966A (en) * | 2014-01-07 | 2014-04-23 | Tcl集团股份有限公司 | Item recommendation method and device |
CN105469263A (en) * | 2014-09-24 | 2016-04-06 | 阿里巴巴集团控股有限公司 | Commodity recommendation method and device |
US20160094501A1 (en) * | 2014-09-26 | 2016-03-31 | Line Corporation | Method, system and recording medium for providing video contents in social platform and file distribution system |
CN105812937A (en) * | 2014-12-30 | 2016-07-27 | Tcl集团股份有限公司 | Television program recommending method and television program recommending device |
CN105989122A (en) * | 2015-02-12 | 2016-10-05 | 广东欧珀移动通信有限公司 | Application recommendation method and device |
CN106227815A (en) * | 2015-07-22 | 2016-12-14 | Tcl集团股份有限公司 | The personalized application program function of a kind of multi-modal clue recommends method and system thereof |
CN105843953A (en) * | 2016-04-12 | 2016-08-10 | 乐视控股(北京)有限公司 | Multimedia recommendation method and device |
CN105930429A (en) * | 2016-04-19 | 2016-09-07 | 乐视控股(北京)有限公司 | Music recommendation method and apparatus |
CN106021485A (en) * | 2016-05-19 | 2016-10-12 | 中国传媒大学 | Multi-element attribute movie data visualization system |
CN106326431A (en) * | 2016-08-25 | 2017-01-11 | 乐视控股(北京)有限公司 | Information recommendation method and device |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110827109A (en) * | 2018-08-14 | 2020-02-21 | 中国移动通信集团浙江有限公司 | Short message marketing method and device |
CN110827109B (en) * | 2018-08-14 | 2022-07-29 | 中国移动通信集团浙江有限公司 | Short message marketing method and device |
CN110152298A (en) * | 2018-11-22 | 2019-08-23 | 腾讯科技(深圳)有限公司 | Realize equipment recommended method, device and the computer equipment of virtual role |
CN109408729A (en) * | 2018-12-05 | 2019-03-01 | 广州市百果园信息技术有限公司 | Material is recommended to determine method, apparatus, storage medium and computer equipment |
CN109408729B (en) * | 2018-12-05 | 2022-02-08 | 广州市百果园信息技术有限公司 | Recommended material determination method and device, storage medium and computer equipment |
WO2020134011A1 (en) * | 2018-12-29 | 2020-07-02 | 北京字节跳动网络技术有限公司 | Method and apparatus for determining display information combination, storage medium, and electronic device |
CN111026934A (en) * | 2019-12-20 | 2020-04-17 | 中科寒武纪科技股份有限公司 | Intelligent recommendation method and related equipment |
CN115329208A (en) * | 2022-10-18 | 2022-11-11 | 平安银行股份有限公司 | Interest fee exemption scheme recommendation method and device, electronic equipment and medium |
Also Published As
Publication number | Publication date |
---|---|
CN108345419B (en) | 2021-06-08 |
WO2018137420A1 (en) | 2018-08-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108345419A (en) | A kind of generation method and device of information recommendation list | |
CN110020094B (en) | Display method and related device for search results | |
CN110704739B (en) | Resource recommendation method and device and computer storage medium | |
CN103823908B (en) | Content recommendation method and server based on user preference | |
CN111782957A (en) | Method and device for recommending content, computing equipment and storage medium | |
CN110008973B (en) | Model training method, method and device for determining target user based on model | |
CN109919311A (en) | The method for generating instruction sequence, the method and apparatus for executing neural network computing | |
WO2021082500A1 (en) | Information recommendation method, apparatus, recommendation server and storage medium | |
CN111459993B (en) | Configuration updating method, device, equipment and storage medium based on behavior analysis | |
CN101859425A (en) | Method and device for providing application list | |
CN110008397B (en) | Recommendation model training method and device | |
CN104731607B (en) | Terminal Lifelong Learning processing method, device and system | |
CN103348342A (en) | Personal content stream based on user-topic profile | |
CN110020175B (en) | Search processing method, processing equipment and system | |
CN108121814B (en) | Search result ranking model generation method and device | |
CN110598120A (en) | Behavior data based financing recommendation method, device and equipment | |
CN110766513A (en) | Information sorting method and device, electronic equipment and readable storage medium | |
CN112104505A (en) | Application recommendation method and device, server and computer-readable storage medium | |
CN115905687A (en) | Cold start-oriented recommendation system and method based on meta-learning graph neural network | |
CN113344647B (en) | Information recommendation method and device | |
CN106997360A (en) | The treating method and apparatus of user behavior data | |
CN109325810B (en) | Recharge conversion improving method, electronic equipment and computer storage medium | |
CN108182600A (en) | A kind of method and system that extending user is determined according to weighted calculation | |
CN111489196A (en) | Prediction method and device based on deep learning network, electronic equipment and medium | |
CN111124209A (en) | Interface display adjustment method and device |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |