CN110324418A - Method and apparatus based on customer relationship transmission service - Google Patents
Method and apparatus based on customer relationship transmission service Download PDFInfo
- Publication number
- CN110324418A CN110324418A CN201910584353.3A CN201910584353A CN110324418A CN 110324418 A CN110324418 A CN 110324418A CN 201910584353 A CN201910584353 A CN 201910584353A CN 110324418 A CN110324418 A CN 110324418A
- Authority
- CN
- China
- Prior art keywords
- user
- couple
- users
- business
- module
- 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
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/22—Indexing; Data structures therefor; Storage structures
- G06F16/2228—Indexing structures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/284—Relational databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0202—Market predictions or forecasting for commercial activities
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L51/00—User-to-user messaging in packet-switching networks, transmitted according to store-and-forward or real-time protocols, e.g. e-mail
- H04L51/52—User-to-user messaging in packet-switching networks, transmitted according to store-and-forward or real-time protocols, e.g. e-mail for supporting social networking services
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/50—Network services
- H04L67/55—Push-based network services
Abstract
Present disclose provides the method and apparatus based on customer relationship transmission service.Specifically, present disclose provides a kind of methods for transmission service, comprising: gather to form a user couple more than first from the first user；For each user couple in more than described first a users couple, prediction model is trained using the behavior label of two users in one or more relationship characteristics of the user couple and the user couple, whether the behavior tag representation of the user user selected the business；More than second a users couple are formed from second user set；For each user couple in more than described second a users couple, predict the user to the probability for selecting the business based on one or more relationship characteristics of the user couple using housebroken prediction model；And selection target user pair set pushes the business from the user couple a more than described second based on the probability.
Description
Technical field
This application involves Internet technical field more particularly to a kind of methods and dress based on customer relationship transmission service
It sets.
Background technique
With the fast development of internet, social networks is at essential a part in for people's lives.With using
The user of platform class application (App) product (for example, Alipay, wechat etc.) is more and more, and platform can push various to user
Business.
In general, can be by using the user characteristic data of each user itself (for example, user when transmission service
Age, gender, educational background, historical behavior data, etc.) predict that the user will click on/buy the probability of certain business (for example, point
Hit rate (CTR), conversion ratio (CVR)), and then determine which user to carry out service propelling to, thus improve the accurate of service propelling
Degree.
But the characteristic of individual consumer is for promoting interaction class business (for example, intimate family, safe preservation etc.) not that
Effectively.It is therefore desirable for a kind of improvement project of more accurately push interaction class business.
Summary of the invention
Present disclose provides a kind of methods for transmission service, comprising:
Gather to form a user couple more than first from the first user；
It is special using one or more relationships of the user couple for each user couple in more than described first a users couple
The behavior label of two users trains prediction model in sign and the user couple, and the behavior tag representation of the user user is
It is no to selected the business；
More than second a users couple are formed from second user set；
For each user couple in more than described second a users couple, the user couple is based on using housebroken prediction model
One or more relationship characteristics predict the user to the probability for selecting the business；And
Based on the probability, selection target user pair set pushes the business from the user couple a more than described second.
Optionally, this method further comprises:
User's transmission service into more than described first a users couple；And
Obtain the behavior label of each user in a user couple more than described first.
Optionally, the acquisition behavior label includes:
If user selects the business, it is determined that the behavior label is the first value；And
If user does not select the business, it is determined that the behavior label is second value.
Optionally, described to gather to form a user more than first to including: from the first user
For each user couple in first user set:
Determine the relationship strength of the user couple；
The relationship strength is compared with first threshold；
If the relationship strength is greater than or equal to the first threshold, by the user to including at more than described first
In user couple；And
If the relationship strength is less than the first threshold, not by the user to including in more than described first a users
Centering.
Optionally, described to form more than second a users to including: from second user set
For each user couple in the second user set:
Determine the relationship strength of the user couple；
The relationship strength is compared with second threshold；
If the relationship strength is greater than or equal to the second threshold, by the user to including at more than described second
In user couple；And
If the relationship strength is less than the second threshold, not by the user to including in more than described second a users
Centering.
Optionally, the relationship strength of the determination user couple includes:
The value of one or more relationship characteristics of the user couple is weighted summation, it is strong with the relationship for obtaining the user couple
Degree.
Optionally, selecting the business includes: click and/or the purchase business.
Optionally, the trained prediction model include: for each user couple in more than described first a users couple, into
One step trains the prediction model using one or more user characteristics of two users in the user couple；And
The prediction probability includes: for each user couple in more than described second a users couple, using housebroken pre-
It surveys model and is based further on one or more user characteristics of the user two users in predict the user described in selection
The probability of business.
Optionally, the relationship characteristic of the user couple include the equipment shared data feature of two users in user couple,
Social networks data characteristics and/or fund relation data feature.
Optionally, the selection target user pair set includes:
More than described second a users are ranked up the probability for selecting the business；And
Select the target user to set according to the sequence.
Optionally, the multiple target users of the selection are to including:
For each user couple in more than described second a users couple, determine that the user is to the probability for selecting the business
It is no to be greater than third threshold value；And
If the user is greater than third threshold value to the probability for selecting the business, by the user to being determined as target user
It is right.
Another aspect of the present disclosure provides a kind of device for transmission service, comprising:
For gathering the module to form a user couple more than first from the first user；
For using one or more relationships of the user couple for each user couple in more than described first a users couple
The behavior label of two users trains the module of prediction model, the behavior tag representation of user in feature and the user couple
Whether the user selected the business；
For forming the module of more than second a users couple from second user set；
For being based on the use using housebroken prediction model for each user couple in more than described second a users couple
One or more relationship characteristics at family pair predict module of the user to the probability for selecting the business；And
For based on the probability, selection target user pair set to push the industry from the user couple a more than described second
The module of business.
Optionally, which further comprises:
Module for from user's transmission service to more than described first a users couple；And
For obtaining the module of the behavior label of each user in a user couple more than described first.
Optionally, the module for obtaining behavior label includes:
If selecting the business for user, it is determined that the behavior label is the module of the first value；And
If not selecting the business for user, it is determined that the behavior label is the module of second value.
Optionally, described to form the module of a user couple more than first for gathering from the first user and include:
For the module for each user in first user set to operation below executing:
Determine the relationship strength of the user couple；
The relationship strength is compared with first threshold；
If the relationship strength is greater than or equal to the first threshold, by the user to including at more than described first
In user couple；And
If the relationship strength is less than the first threshold, not by the user to including in more than described first a users
Centering.
Optionally, include: for the module for forming more than second a users couple from second user set
For the module for each user in the second user set to operation below executing:
Determine the relationship strength of the user couple；
The relationship strength is compared with second threshold；
If the relationship strength is greater than or equal to the second threshold, by the user to including at more than described second
In user couple；And
If the relationship strength is less than the second threshold, not by the user to including in more than described second a users
Centering.
Optionally, the module for determining the relationship strength of the user couple includes:
For the value of one or more relationship characteristics of the user couple to be weighted summation, to obtain the pass of the user couple
It is the module of intensity.
Optionally, selecting the business includes: click and/or the purchase business.
Optionally, the module for being used to train prediction model includes: for in more than described first a users couple
Each user couple, further use one or more user characteristics of two in the user couple users to train the prediction
The module of model；And
The module for prediction probability includes: each user couple for being directed in more than described second a users couple,
One or more user characteristics of two users in the user couple are based further on using housebroken prediction model to predict this
Module of the user to the probability for selecting the business.
Optionally, the relationship characteristic of the user couple include the equipment shared data feature of two users in user couple,
Social networks data characteristics and/or fund relation data feature.
Optionally, the module for selection target user pair set includes:
Module for probability of more than the described second a users to the selection business to be ranked up；And
For selecting the target user to the module of set according to the sequence.
Optionally, the module for being used to select multiple target users couple includes:
For determining the user to the general of the selection business for each user couple in more than described second a users couple
Whether rate is greater than the module of third threshold value；And
If being greater than third threshold value to the probability for selecting the business for the user, by the user to being determined as target
The module of user couple.
On the one hand having for the disclosure provides a kind of device for transmission service, comprising:
Processor；And
It is arranged to the memory of storage computer executable instructions, the executable instruction makes the place when executed
It manages device and executes following operation:
Gather to form a user couple more than first from the first user；
It is special using one or more relationships of the user couple for each user couple in more than described first a users couple
The behavior label of two users trains prediction model in sign and the user couple, and the behavior tag representation of the user user is
It is no to selected the business；
More than second a users couple are formed from second user set；
For each user couple in more than described second a users couple, the user couple is based on using housebroken prediction model
One or more relationship characteristics predict the user to the probability for selecting the business；And
Based on the probability, selection target user pair set pushes the business from the user couple a more than described second.
Detailed description of the invention
Fig. 1 is according to various aspects of the disclosure for the system diagram based on customer relationship transmission service.
Fig. 2 be include user 1-4 user set in customer relationship diagram.
Fig. 3 be according to all aspects of this disclosure based on user to relationship come the diagram of the method for transmission service.
Fig. 4 be according to all aspects of this disclosure based on user to relationship come the flow chart of the method for transmission service.
Fig. 5 is to be used for based on user to relationship according to all aspects of this disclosure come the diagram of the device of transmission service.
Specific embodiment
For the above objects, features, and advantages of the disclosure can be clearer and more comprehensible, below in conjunction with attached drawing to the tool of the disclosure
Body embodiment elaborates.
Many details are explained in the following description in order to fully understand the disclosure, but the disclosure can be with
It is different from other way described herein using other and implements, therefore the disclosure is by the limit of following public specific embodiment
System.
It routinely, can be according to the user characteristic data of user itself (for example, the age of user, star in service propelling
Seat, educational background, region, income etc.) predict that user will click on/buy the probability of certain business, but user characteristic data for
Click/buying rate the prediction for interacting class business is not accurate enough.
Interacting class business is the form of service emerged in large numbers recently, can be related to the interaction behaviour between two or more users
Make, for example, it is intimate pay, safe preservation etc..It is that another account is paid that intimate pair, which can permit using an account, example
Such as, shopping payment, transfer accounts etc..Two interlock accounts can be respectively defined as guarding account and be guarded account by safe preservation,
If having unusual condition by account is guarded, can be sounded an alarm to account is guarded.
Interaction class business often relates to the relation data between multiple users.For example, intimate two paid with safe preservation
Relationship between the user of interlock account is often kinsfolk (for example, spouse, parent, children etc.) or close friend.
Therefore when pushing interaction class business to user, the relationship between two or more users is accounted for can be improved a little
Hit/buying rate prediction accuracy.
The present disclosure contemplates that the above-mentioned characteristic of interaction class business, in order to make up the deficiency of individual consumer's feature, in prediction industry
Using user to relationship characteristic during click/purchase probability of business.User is two user (users of characterization to relationship characteristic
It is right) between relationship feature.
For example, user may include following three classes to relationship characteristic:
One, intermediary relationships feature, it may include two users are about medium (for example, equipment, gateway, telephone number, mailbox
Deng) shared data (for example, share number, share duration etc.)；
Two, social networks feature, it may include the mobile phone of two users/telephone relation information, social software is (for example, Taobao
Wang Wang, wechat, microblogging etc.) operation information (for example, addition good friend, chat number, hop count, comment number), etc.；
Three, fund relationship characteristic, it may include paying out between two users, transfer data (for example, number, amount of money etc.).
It is enumerated above several examples of the relationship characteristic of user couple, but skilled artisans will appreciate that, user's couple
Relationship characteristic is not limited only to above example, as long as the data for the close relation degree being able to reflect between different user are all in this public affairs
In the conception opened.
Further, user is often operated using account, therefore is interchangeable in this paper term " user " and " account "
What ground used.
Fig. 1 is the diagram for the system 100 based on customer relationship transmission service according to various aspects of the disclosure.
As shown in Figure 1, system 100 may include multiple 101 1-N of terminal, server 102 and database 103.Multiple terminals
It can be communicated by wired or wireless connection between 101-1-N, server 102 and database 103.
Terminal 101 can be the device with network connecting function, for example, smart phone, laptop, plate are electric
Brain, desktop computer etc..Application program (for example, Alipay, wechat etc.) can be run in terminal 101.
Server 102 can be a server, be also possible to include multiple servers cluster of servers.Server
102 can provide various business services for multiple terminal 101-1-N.
Each terminal 101 can carry out various operations, including interactive operation by server 102.For example, these are operated
It may include sending mail, voice communication, being paid out using app chat, transferring accounts between account, account etc..
Further, server 102 can collect data from each terminal 101, for example, two users between (user to)
The personal feature of relationship characteristic (relationship characteristic of user couple, referred to as relationship characteristic) and/or user (are referred to herein as user spy
Sign).It is closed for example, server 102 can be collected in the operation log and/or operation information (for example, operation requests) of terminal 101
It is feature and user characteristics.
Relationship characteristic may include intermediary relationships feature, social networks feature and fund relationship characteristic as described above.User
Feature may include age of user itself, constellation, educational background, income, historical behavior etc..
Server 102 can store the relationship characteristic being collected into and user characteristics into database 103 for subsequent use.
Database 103 can store the relationship characteristic and optional user characteristics of the behavior label data of user, user couple.
Server 102 further may include for predicting that user selects the prediction model of the probability of business (for example, there is supervision
Model).Training set can be used to train this to have a monitor model in server 102, and training set may include user's set (including multiple use
Family to) in user behavior label (indicate user whether selected business, such as, if click, purchase business) and should
The relationship characteristic of multiple users couple and optional user characteristics train this to have monitor model；May then use that housebroken has
Monitor model predicts the behavioral data of other users, for example, probability of the user to selection (click or purchase) business.
Although note that in Fig. 1, server 102 and database 103 are shown separately, and database 103 can also be received
Enter in server 102.
Fig. 2 is the diagram for including the user in user's set of four users (that is, user 1-4) to relationship.
User in user's set may include the relationship between any two user in user set to relationship.Such as figure
Described in 2, the user of user including user 1, user 2, user 3 and user 4 set is to relationship can include: (user 1 to 12 by user
With user 2), user to 13 (user 1 and users 3), user to 14 (user 1 and users 4), user to 23 (user 2 and users
3), user is to 24 (user 2 and users 4) and user to the relationship between 34 (user 3 and users 4).
User including n user, which gathers, may includeA user couple.
The mark that two users can be used in the relationship characteristic of user couple is stored indexed by database 103.Table 1 shows
The user for having gone out user's set including n user is to an example of the storage organization of feature, wherein the mark of each user
It is indicated using digital 1-n.Skilled artisans will appreciate that the mode of other identity users pair is also in the conception of the disclosure
In.Each user can have one or more relationship characteristics to relationship.
Table 1
User may include shared data (example of two users for equipment, gateway, telephone number, mailbox to relationship characteristic
Such as, it shares number, share duration)；The behavior of good friend, chat number, forwarding time is mutually added in two users in social software
Number, comment number etc.；Treasury trade data between two users, for example, paying out, number of transferring accounts, the amount of money etc..
Fig. 3 be according to all aspects of this disclosure based on user to relationship characteristic come the diagram of the method for transmission service.
In the disclosure based on user to relationship characteristic come in the method for transmission service, server 102 is first on a small scale
Dispensing business, for example, randomly (the first user set is shown as in Fig. 3, for example, several hundred or several to the subset of associated user
Thousand users) business is launched, and the behavior label of each user in user's subset is obtained, whether behavior tag representation user selects
The business is selected, such as, if click the business, whether buy the business etc..Subsequent server 102 can be obtained from database 103
The user of user's subset, and (and can using acquired relationship characteristic to relationship characteristic (and optional user characteristics)
Optional user characteristics) and the behavior label of user trained monitor model.After training has monitor model, server
102 can be used it is housebroken have monitor model predict user couple behavioral data (user to selection business probability, for example,
The probability of click-to-call service, the probability for buying business).
Training has the behavior label of monitor model and the use of the behavioral data for having monitor model to predict is same type.Example
Such as, if system will predict the behavior mark that the user in second user set obtains the probability of click-to-call service, server 102
Whether the user that label can be in the first user set has clicked the business；If system will predict the use in second user set
To the probability of purchase business, then the behavior label that server 102 obtains is whether the user in the first user set buys this at family
Business, and so on.
In the following description, it is explained with behavior label and behavioral data for purchase business, but art technology
Personnel will be appreciated that, other types of behavior label and behavioral data are also in the conception of the disclosure.
As shown in figure 3, server 102 can be to user's transmission service in the first user set in step 301.
First user set can be the son of associated user's set (for example, potential user's set of purchase business) of business
Collection.The subset that the associated user that server 102 can be randomly chosen business combines is gathered as the first user.Further, it takes
Business device 102 can be gathered to form a user couple more than first from the first user, and to more than first a users to transmission service.
Specifically, the user in the first user set can be formed two-by-two a user couple by server 102, to be formed
A user couple more than first, for example, as shown for example in figure 2.The first user set including n user can be formedA use
Family pair.
Optionally, server 102 can the first user gather in whole users couple in filter out relationship strength compared with
Multiple users of high (for example, being higher than threshold value) to as more than first a users to carrying out service propelling.
For example, for each user couple formed in the first user set, it can be by multiple relationship characteristics of the user couple
Value be weighted summation with determine its relationship strength S.If relationship strength S is higher than a threshold value (the first intensity threshold), can
By the user to including in user couple a more than first with transmission service, for the operation of subsequent model training；If closed
It is intensity S lower than the first intensity threshold, then it can not be by the user to including in user couple a more than first.
If the relationship strength of user couple is lower, illustrate the relationship defective tightness of two users of user couple, for interaction
The reference significance of the data prediction of class business is little.
More than first a users are screened to the calculation amount that can reduce training pattern by user's intensity, and the business of saving pushes away
The communication resource sent.
In step 302, server 102 can obtain more than first behavior label of each user about business in a user couple.
The behavior label of user can characterize whether the user selects (for example, click or buy) business.
For example, server 102 can determine that the behavior label of the user is the first value (example if user buys the business
Such as, 1)；If user does not buy the business, server 102 can determine that the behavior label of the user is second value (for example, 0).
Here value is only example, and other values are also possible.
After server 102 obtains the behavior label of user, it can store it in database 103 for subsequent use.
In step 303, server 102 can obtain the relationship characteristic of more than first a users couple from database 103.
For each user couple in more than first a users couple, server 102 can search the user in database 103
Pair one or more relationship characteristics relevant to the business and two users of the user couple respective one or more use
Family feature.
User to relationship characteristic may include two users for equipment, gateway, telephone number shared data (for example, altogether
With number, share duration)；The behavior of good friend is mutually added in social software, chat number, hop count, comments by two users
By number etc.；Treasury trade data between two users, for example, paying out, number, the amount of money transferred accounts etc..
User characteristics may include age of user, gender, constellation, region, membership information, historical behavior label (for example, closing
In historic click-through rate, the conversion ratio etc. of related service), etc..
Although step 301 and 302 is before step 303 in Fig. 3, step 303 can also be before step 301 or 302
It executes.
Server 102 can also obtain the user characteristics of the user more than first in a user couple to instruct for following model
Practice.
In step 304, the row of the user in more than first a users couple that step 302 obtains is can be used in server 102
The relationship characteristic of the user couple obtained for label and in step 303 trains prediction model.
Preferably, server 102 can further use the user characteristics of the user more than first in a user couple to train
Model.
Server 102 can select the relationship characteristic of user couple and optional user special according to the characteristic of business
Sign, is quantified and forms feature vector.Server 102 can be by the behavior mark of two users in this feature vector sum user couple
Label carry out training pattern as input.
For example, intimate pay two users related generally in kinsfolk.In this case, user can to relationship characteristic
Gateway, the number of computer equipment and/or duration are shared with the account for including two users, this is because kinsfolk is often
It is operated in the identical gateway online of residential usage or using identical computer equipment logon account；User is to relationship spy
Sign can also include chat number, transfer accounts record (transfer accounts number, the amount of money) of the account of two users in chat software；User
It further may include that the account of two users shares the information of telephone number or mailbox to relationship characteristic, for example, two accounts can
Identical telephone number or mailbox can be registered on platform (for example, Alipay, wechat).
Preferably, intimate to repay the user characteristics that can be related to user itself, for example, two users respective age, gender,
Region etc..For example, if the gender of two users is respectively male and female and age difference is relatively close (for example, less than ten years old),
Two users are that the probability of spouse is larger；If the age difference 20 to three of two users ten years old, the relationship of two users
It may be the relationship of parent and child, it is higher using the probability intimately paid.As another example, user characteristics may also include payment
Ability, it is lower using the probability intimately paid if the ability to pay of two users is all very low；If at least one in two users
The ability to pay of person is higher, then higher using the probability intimately paid.As another example, user characteristics further may include two
The account liveness of user, for example, making if the account liveness of two users is all higher (for example, it is longer to enliven number of days)
It is also higher with the probability intimately paid.
User is illustrated for intimately paying above to the example of relationship characteristic and user characteristics, but the relationship of the disclosure is special
User characteristics of seeking peace are not limited to this.Skilled artisans will appreciate that other relationship characteristics and user characteristics also can be used, and
And different relationship characteristic and user characteristics can be considered in different business.
For the first user set in each user couple, server 102 can be used the relationship characteristic of the user couple with
And optional user characteristics form feature vector, and this feature vector and the input of the behavior label of two users are had supervision mould
Type, so that this be trained to have monitor model.
E.g., including the feature vector of the user couple of user i and user j can be expressed as:
fij=[P1,P2,…,Pa,I1,I2,…,Ib,J1,J2,…,Jc,],
Wherein P1,P2,…,PaFor relationship characteristic value, I1,I2,…,IbFor the user characteristics value of user i, J1,J2,…,JcFor
The user characteristics value of user j.
As an example, if the relationship characteristic of user couple includes P1=chat number (159), P2=share computer
Duration (512 (hour)), P3=transfer amounts (15600 (member))；The user characteristics of user i include I1=gender (1/ female), I2=
Age (30), the user characteristics of user j include J1=gender (0/ male), J2=age (35), then feature vector fij=[159,
512,15600,1,30,0,30]。
The feature vector of user couple has been input to by server 102 together with the behavior label of two users in user couple
Monitor model is trained.
For example, the input of model can be { fij, Li,Lj}。
Wherein LiFor the behavior label of user i, and LjFor the behavior label of user j.For example, if user i has purchased industry
It is engaged in, then LiIt can be 1, if user i does not buy business, LiIt can be 0；If user j has purchased business, LjCan be
1, if user j does not buy business, LjIt can be 0.
Above example carrys out training pattern using relationship characteristic and user characteristics, but can also train using only relationship characteristic
Model.
In step 305, server 102 obtains the use of more than second a users couple in second user set from database 103
The relationship characteristic at family pair.
Second user set can be associated user's set of business.Server 102 can be predicted in second user set
The behavioral data of each user, and according to behavioral data, selection target user gathers come transmission service from second user set.
User in second user set can be formed a user couple by server 102 two-by-two, to form more than second
A user couple.
Optionally, server 102 can be filtered out in whole users couple in second user set relationship strength compared with
Multiple users of high (for example, being higher than the second intensity threshold) are to as more than second a users couple.
For example, for each user couple formed in second user set, it can be by multiple relationship characteristics of the user couple
Value be weighted summation with determine its relationship strength S.If relationship strength S is higher than the second intensity threshold, can be by the user
To being included in more than second in a user couple to carry out subsequent predicted operation；If relationship strength S is lower than the second intensity threshold,
It can not be by the user to including in user couple a more than second.
For example, if the user including user i and user j is to relationship characteristic value P1,P2,…,Pa, then can be as follows
Calculate the relationship strength of the user couple:
S=ω1P1+ω2P2+…+ωaPa。
Wherein 0≤ωi≤ 1, ωiValue can select according to actual needs.
If relationship strength S is lower than a threshold value, illustrate the relationship defective tightness of two users, is less likely purchase interaction class
Business then can reject two users in the user couple from associated user's set, without subsequent predicted operation.Change speech
It, server 102 can only by associated user gather in relationship strength S be higher than two users in the user couple of threshold value and be included in the
In two users set, it is possible thereby to reduce the calculation amount of model prediction.
In step 306, server 102 is predicted using the relationship characteristic of a user couple more than second obtained in step 305
The user is to the behavioral data about business.
Specifically, for each user couple, it may for example comprise the user of user i and user j, can be by users couple to i-j
The relationship characteristic of i-j inputs housebroken prediction model to predict user to i-j purchase/click-to-call service probability.
Optionally, also the user characteristics of the user characteristics of user i and user j can be inputted warp together with relationship characteristic
Trained model is predicted.
Relationship characteristic and optional user characteristics included in the input of prediction model can be with steps in step 306
It is corresponding for included relationship characteristic in the input of training pattern and optional user characteristics in rapid 304.
In step 307, server 102 can determine the target user in second user set to set.
For example, the behavioral data of user couple can be ranked up by server 102, select user in the top to as mesh
Mark user couple.
As another example, a threshold value also can be set in server 102, is somebody's turn to do if the predictive behavior data of user couple are higher than
Threshold value, then by the user to being determined as target user couple.
In step 308, server 102 can be to target user to user's transmission service in set.
It note that describe in the embodiment above and obtain user in step 303 and 305 to relationship characteristic and user spy
Sign comes training pattern and predictive behavior data, but can also instruct using only user to relationship characteristic, without the use of user characteristics
Practice model and predictive behavior data.
Fig. 4 be according to all aspects of this disclosure based on user to relationship come the flow chart of the method for transmission service.
In step 402, can gather to form a user couple more than first from the first user.
First user set can be the son of associated user's set (for example, potential user's set of purchase business) of business
Collection.Whether the user in the relationship characteristic of a user couple more than first and more than first a users couple of instruction selects the behavior label of business
It can be used to training pattern.
On the one hand, all users in the first user being gathered form user to as more than first a users two-by-two
It is right.
On the other hand, can the relationship strength based on user couple to all users in the first user set to screening
To form more than first a users couple.It is somebody's turn to do for example, summation can be weighted to the value of multiple relationship characteristics of user couple with determining
The relationship strength S of user couple.It, can be by the user to from more than first a users if relationship strength S is lower than the first intensity threshold
Centering is rejected, without subsequent model training.
In step 404, the relationship characteristic of a user couple more than first can be used and behavior label carrys out training pattern.
Specifically, can then obtain the behavior of each user to user's transmission service in more than first a users couple
Whether label, the behavior tag characterization user select the business, such as, if click or buy the business.
The relationship characteristic of user couple may include two users about medium shared data feature (for example, two users couple
In equipment, gateway, telephone number shared data (for example, share number, share duration))；The social networks number of two users
According to feature (for example, the behavior of good friend, chat number, hop count, comment number is mutually added in two users in social software
Deng)；Fund relationship characteristic (for example, the treasury trade data between two users, for example, number, the amount of money paying out, transfer accounts
Deng).
The relationship characteristic of a user couple more than first and behavior label can be used to train prediction model.
In step 406, more than second a users couple can be formed from second user set.
Second user set can be associated user's set of business.
It on the one hand, can be by the way that user in the associated user of business set be formed user to being formed more than second two-by-two
A user couple.
On the other hand, can the relationship strength based on user couple to all users in second user set to screening
To form more than second a users couple.For example, multiple relationship characteristics of user couple can be weighted with summation to determine the user
Pair relationship strength S.If relationship strength S is lower than the second intensity threshold, can be by the user to from user couple a more than second
It rejects, without subsequent predicted operation.
In step 408, housebroken prediction model can be used based on the relationship characteristic of more than second a users couple to predict
Probability of the user to selection business.
For example, relationship characteristic of the user to ij can be inputted through instructing for the user including user i and user j to ij
Experienced model selects the probability (for example, probability that user i and user j select business) of business to predict user to ij.
Preferably, user can also be predicted to the probability of ij selection business using the user characteristics of user i and user j.
In step 410, service propelling is carried out come selection target user pair set based on the probability obtained in step 408.
For example, probability of the user to selection business can be ranked up, select multiple users in the top to carrying out group
At target user to set.
As another example, a threshold value can be set, it, should if user is higher than the threshold value to the probability of selection business
User is to being determined as target user couple.
It then can be to target user to transmission service.
Optionally, it can also use user's itself in the predicted operation of the training pattern of step 404 and step 408
User characteristics.
Fig. 5 is to be used for based on user to relationship according to all aspects of this disclosure come the process schematic of transmission service.
501, the feature of a user couple more than available first.
A user more than first to be formed to that can gather from the first user.First user set can be associated user's collection of business
Close the subset of (for example, potential user's set of purchase business).For example, the first user set can be randomly selected user's
The data (for example, the behavior label of user to relationship characteristic, user) of collection, the first user set can be used to training pattern.It can
The user in the first user set is formed a user couple two-by-two, to form more than first a users couple.It optionally, can be with
The weaker user couple of some relationships is weeded out from user couple a more than first according to relationship strength, to reduce the meter of training pattern
Calculation amount.
The feature of a user couple more than first may include one or more relationships of each user couple in a user couple more than first
Feature.Optionally, the feature of a user couple more than first may also include more than first one or more of each user in a user couple
A user characteristics.
502, the behavior label of a user couple more than available first.
For example, user's transmission service that can in advance into more than first a users couple, and whether selected according to each user
(click or purchase) business generates the behavior label of the user.
For example, can determine that the behavior label of the user is the first value if user selects the business；If user is unselected
The business is selected, then can determine that the behavior label of the user is second value.
The behavior label of a user couple more than the feature of a user couple more than the first of 501 and the first of 502 can be used to instruct
Practice model 504.
503, the feature of a user couple more than available second.
A user more than second from second user set to can generate.Second user set can be associated user's collection of business
It closes (for example, the user of potential purchase business gathers).User in second user set can be formed to a user couple two-by-two,
To form more than second a users couple.Optionally, some passes can be weeded out from user couple a more than second according to relationship strength
It is weaker user couple, to reduce the calculation amount of prediction.
The feature of a user couple more than second may include one or more relationships of each user couple in a user couple more than second
Feature.Optionally, the feature of a user couple more than second may also include one or more of each user more than second in a user couple
A user characteristics.
The feature input model 504 of more than second a users couple can be predicted to more than second each user couple in a user couple
Behavioral data 505, for example, user (clicks or buy) probability of the business to selection.
506, target user can be determined according to behavioral data 505.
For example, the behavioral data of user couple can be ranked up, select user in the top to as target user couple.
Alternatively, the predictive behavior data of user couple can be compared with a threshold value, if the prediction row of user couple
It is that data are higher than the threshold value, then by the user to being determined as target user couple.
507, to target user to set transmission service.
Specifically, to target user to each user's transmission service in set.
The disclosure, using the relationship characteristic between two users, is examined during clicking rate/conversion ratio of the business of prediction
The close relation degree of two users is considered, to effectively improve the pushing efficiency of interaction class business.
Preferably, the disclosure is chosen the training set of training pattern and is determined the mistake of target user couple using housebroken model
Cheng Zhong screens user couple by relationship strength, it is possible thereby to reduce calculation amount.
Claim can be implemented or fall in without representing by describing example arrangement herein in conjunction with the explanation that attached drawing illustrates
In the range of all examples.Term as used herein " exemplary " means " being used as example, example or explanation ", and simultaneously unexpectedly
Refer to " being better than " or " surpassing other examples ".This detailed description includes detail to provide the understanding to described technology.So
And these technologies can be practiced without these specific details.In some instances, it well-known structure and sets
It is standby to be shown in block diagram form to avoid fuzzy described exemplary concept.
In the accompanying drawings, similar assembly or feature can appended drawing references having the same.In addition, the various components of same type can
It is distinguish by the second label distinguished followed by dash line and between similar assembly in appended drawing reference.If
The first appended drawing reference is used only in the description, then the description can be applied to the similar assembly of the first appended drawing reference having the same
Any one of component regardless of the second appended drawing reference how.
It can be described herein with being designed to carry out in conjunction with the various illustrative frames and module of open description herein
The general processor of function, DSP, ASIC, FPGA or other programmable logic device, discrete door or transistor logic, point
Vertical hardware component, or any combination thereof realize or execute.General processor can be microprocessor, but in alternative
In, processor can be any conventional processor, controller, microcontroller or state machine.Processor can also be implemented as counting
The combination of equipment is calculated (for example, DSP and the combination of microprocessor, multi-microprocessor, the one or more cooperateed with DSP core
Microprocessor or any other such configuration).
Function described herein can hardware, the software executed by processor, firmware, or any combination thereof in it is real
It is existing.If realized in the software executed by processor, each function can be used as one or more instruction or code is stored in
It is transmitted on computer-readable medium or by it.Other examples and realization fall in the disclosure and scope of the appended claims
It is interior.For example, function described above can be used the software executed by processor, hardware, firmware, connect firmly due to the essence of software
Line or any combination thereof is realized.It realizes that the feature of function can also be physically located in various positions, including is distributed so that function
Each section of energy is realized in different physical locations.In addition, being arranged as used in (including in claim) herein in project
It lifts and is used in (for example, being enumerated with the project with the wording of such as one or more of at least one of " " or " " etc)
"or" instruction inclusive enumerate so that such as at least one of A, B or C enumerate mean A or B or C or AB or AC or
BC or ABC (that is, A and B and C).Equally, as it is used herein, phrase " being based on " is not to be read as citation sealing condition collection.
Illustrative steps for example, be described as " based on condition A " can model based on both condition A and condition B without departing from the disclosure
It encloses.In other words, as it is used herein, phrase " being based on " should be solved in a manner of identical with phrase " being based at least partially on "
It reads.
Computer-readable medium includes both non-transitory, computer storage medium and communication media comprising facilitates computer
Any medium that program shifts from one place to another.Non-transitory storage media, which can be, to be accessed by a general purpose or special purpose computer
Any usable medium.Non-limiting as example, non-transient computer-readable media may include that RAM, ROM, electric erasable can
Program read-only memory (EEPROM), compact disk (CD) ROM or other optical disc storages, disk storage or other magnetic storage apparatus,
Or it can be used to carry or store instruction or the expectation program code means of data structure form and can be by general or specialized calculating
Machine or any other non-transitory media of general or specialized processor access.Any connection is also properly termed computer
Readable medium.For example, if software is using coaxial cable, fiber optic cables, twisted pair, digital subscriber line (DSL) or such as red
Outside, the wireless technology of radio and microwave etc is transmitted from web site, server or other remote sources, then should
Coaxial cable, fiber optic cables, twisted pair, digital subscriber line (DSL) or such as infrared, radio and microwave etc it is wireless
Technology is just included among the definition of medium.As used herein disk (disk) and dish (disc) include CD, laser disc, light
Dish, digital universal dish (DVD), floppy disk and blu-ray disc, which disk usually magnetically reproduce data and dish with laser come optically again
Existing data.Combination of the above media is also included in the range of computer-readable medium.
There is provided description herein is in order to enable those skilled in the art can make or use the disclosure.To the disclosure
Various modifications will be apparent those skilled in the art, and the generic principles being defined herein can be applied to it
He deforms without departing from the scope of the present disclosure.The disclosure is not defined to examples described herein and design as a result, and
It is that the widest scope consistent with principles disclosed herein and novel feature should be awarded.
Claims (23)
1. a kind of method for transmission service, comprising:
Gather to form a user couple more than first from the first user；
For each user couple in more than described first a users couple, using one or more relationship characteristics of the user couple, with
And the behavior label of two users trains prediction model in the user couple, whether the behavior tag representation of the user user selected
Select the business；
More than second a users couple are formed from second user set；
For each user couple in more than described second a users couple, using housebroken prediction model based on the user couple one
A or multiple relationship characteristics predict the user to the probability for selecting the business；And
Based on the probability, selection target user pair set pushes the business from the user couple a more than described second.
2. the method as described in claim 1, which is characterized in that further comprise:
User's transmission service into more than described first a users couple；And
Obtain the behavior label of each user in a user couple more than described first.
3. method according to claim 2, which is characterized in that the acquisition behavior label includes:
If user selects the business, it is determined that the behavior label is the first value；And
If user does not select the business, it is determined that the behavior label is second value.
4. the method as described in claim 1, which is characterized in that described to gather to form a user more than first to packet from the first user
It includes:
For each user couple in first user set:
Determine the relationship strength of the user couple；
The relationship strength is compared with first threshold；
If the relationship strength is greater than or equal to the first threshold, by the user to including in more than described first a users
Centering；And
If the relationship strength is less than the first threshold, not by the user to including in more than described first a users couple
In.
5. the method as described in claim 1, which is characterized in that described to form more than second a users to packet from second user set
It includes:
For each user couple in the second user set:
Determine the relationship strength of the user couple；
The relationship strength is compared with second threshold；
If the relationship strength is greater than or equal to the second threshold, by the user to including in more than described second a users
Centering；And
If the relationship strength is less than the second threshold, not by the user to including in more than described second a users couple
In.
6. method as described in claim 4 or 5, which is characterized in that the relationship strength of the determination user couple includes:
The value of one or more relationship characteristics of the user couple is weighted summation, to obtain the relationship strength of the user couple.
7. the method as described in claim 1, which is characterized in that selecting the business includes: click and/or the purchase industry
Business.
8. the method as described in claim 1, which is characterized in that
The trained prediction model includes: to further use the use for each user couple in more than described first a users couple
One or more user characteristics of two users of family centering train the prediction model；And
The prediction probability includes: to use housebroken prediction mould for each user couple in more than described second a users couple
Type is based further on one or more user characteristics of two users in the user couple to predict the user to the selection business
Probability.
9. the method as described in claim 1, which is characterized in that the relationship characteristic of the user couple includes two use in user couple
Equipment shared data feature, social networks data characteristics and/or the fund relation data feature at family.
10. the method as described in claim 1, which is characterized in that the selection target user pair set includes:
More than described second a users are ranked up the probability for selecting the business；And
Select the target user to set according to the sequence.
11. the method as described in claim 1, which is characterized in that the multiple target users of the selection are to including:
For each user couple in more than described second a users couple, determine whether the user is big to the probability for selecting the business
In third threshold value；And
If the user is greater than third threshold value to the probability for selecting the business, by the user to being determined as target user couple.
12. a kind of device for transmission service, comprising:
For gathering the module to form a user couple more than first from the first user；
For using one or more relationships spy of the user couple for each user couple in more than described first a users couple
The behavior label of two users trains the module of prediction model in sign and the user couple, and the behavior tag representation of user should
Whether user selected the business；
For forming the module of more than second a users couple from second user set；
For being based on the user couple using housebroken prediction model for each user couple in more than described second a users couple
One or more relationship characteristics predict module of the user to the probability for selecting the business；And
For based on the probability, selection target user pair set to push the business from the user couple a more than described second
Module.
13. device as claimed in claim 12, which is characterized in that further comprise:
Module for from user's transmission service to more than described first a users couple；And
For obtaining the module of the behavior label of each user in a user couple more than described first.
14. device as claimed in claim 13, which is characterized in that the module for obtaining behavior label includes:
If selecting the business for user, it is determined that the behavior label is the module of the first value；And
If not selecting the business for user, it is determined that the behavior label is the module of second value.
15. device as claimed in claim 12, which is characterized in that described to form more than first use for gathering from the first user
The module at family pair includes:
For the module for each user in first user set to operation below executing:
Determine the relationship strength of the user couple；
The relationship strength is compared with first threshold；
If the relationship strength is greater than or equal to the first threshold, by the user to including in more than described first a users
Centering；And
If the relationship strength is less than the first threshold, not by the user to including in more than described first a users couple
In.
16. device as claimed in claim 12, which is characterized in that form more than second use from second user set for described
The module at family pair includes:
For the module for each user in the second user set to operation below executing:
Determine the relationship strength of the user couple；
The relationship strength is compared with second threshold；
If the relationship strength is greater than or equal to the second threshold, by the user to including in more than described second a users
Centering；And
If the relationship strength is less than the second threshold, not by the user to including in more than described second a users couple
In.
17. the device as described in claim 15 or 16, which is characterized in that described for determining the relationship strength of the user couple
Module includes:
It is strong with the relationship for obtaining the user couple for the value of one or more relationship characteristics of the user couple to be weighted summation
The module of degree.
18. method as claimed in claim 12, which is characterized in that selecting the business includes: click and/or the purchase industry
Business.
19. device as claimed in claim 12, which is characterized in that
The module for being used to train prediction model includes: each user couple for being directed in more than described first a users couple,
One or more user characteristics of two in the user couple users are further used to train the module of the prediction model；And
And
The module for prediction probability includes: for using for each user couple in more than described second a users couple
Housebroken prediction model is based further on one or more user characteristics of two users in the user couple to predict the user
Module to the probability for selecting the business.
20. device as claimed in claim 12, which is characterized in that the relationship characteristic of the user couple includes two in user couple
Equipment shared data feature, social networks data characteristics and/or the fund relation data feature of user.
21. device as claimed in claim 12, which is characterized in that the module packet for selection target user pair set
It includes:
Module for probability of more than the described second a users to the selection business to be ranked up；And
For selecting the target user to the module of set according to the sequence.
22. device as claimed in claim 12, which is characterized in that described for selecting the module packet of multiple target users couple
It includes:
For determining that the user is to the probability for selecting the business for each user couple in more than described second a users couple
The no module greater than third threshold value；And
If being greater than third threshold value to the probability for selecting the business for the user, by the user to being determined as target user
Pair module.
23. a kind of device for transmission service, comprising:
Processor；And
It is arranged to the memory of storage computer executable instructions, the executable instruction makes the processor when executed
Execute following operation:
Gather to form a user couple more than first from the first user；
For each user couple in more than described first a users couple, using one or more relationship characteristics of the user couple, with
And the behavior label of two users trains prediction model in the user couple, whether the behavior tag representation of the user user selected
Select the business；
More than second a users couple are formed from second user set；
For each user couple in more than described second a users couple, using housebroken prediction model based on the user couple one
A or multiple relationship characteristics predict the user to the probability for selecting the business；And
Based on the probability, selection target user pair set pushes the business from the user couple a more than described second.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910584353.3A CN110324418B (en) | 2019-07-01 | 2019-07-01 | Method and device for pushing service based on user relationship |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910584353.3A CN110324418B (en) | 2019-07-01 | 2019-07-01 | Method and device for pushing service based on user relationship |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110324418A true CN110324418A (en) | 2019-10-11 |
| CN110324418B CN110324418B (en) | 2022-09-20 |
Family
ID=68122142
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910584353.3A Active CN110324418B (en) | 2019-07-01 | 2019-07-01 | Method and device for pushing service based on user relationship |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110324418B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111340605A (en) * | 2020-05-22 | 2020-06-26 | 支付宝(杭州)信息技术有限公司 | Method and device for training user behavior prediction model and user behavior prediction |
| CN112243021A (en) * | 2020-05-25 | 2021-01-19 | 北京沃东天骏信息技术有限公司 | Information pushing method, device, equipment and computer readable storage medium |
Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103150663A (en) * | 2013-02-18 | 2013-06-12 | 亿赞普(北京)科技有限公司 | Method and device for placing network placement data |
| CN103955545A (en) * | 2014-05-22 | 2014-07-30 | 成都品果科技有限公司 | Personalized social network influence identifying method |
| US20160241605A1 (en) * | 2015-02-17 | 2016-08-18 | Google Inc. | Second-screen control automatic pairing using push notifications |
| CN106570014A (en) * | 2015-10-09 | 2017-04-19 | 阿里巴巴集团控股有限公司 | Method and device for determining home attribute information of user |
| CN107103057A (en) * | 2017-04-13 | 2017-08-29 | 腾讯科技（深圳）有限公司 | A kind of resource supplying method and device |
| CN107464141A (en) * | 2017-08-07 | 2017-12-12 | 北京京东尚科信息技术有限公司 | For the method, apparatus of information popularization, electronic equipment and computer-readable medium |
| CN107895277A (en) * | 2017-09-30 | 2018-04-10 | 平安科技（深圳）有限公司 | Method, electronic installation and the medium of push loan advertisement in the application |
| US20180150464A1 (en) * | 2016-11-30 | 2018-05-31 | Facebook, Inc. | Multi-stage Filtering for Recommended User Connections on Online Social Networks |
| CN108460590A (en) * | 2018-02-06 | 2018-08-28 | 北京三快在线科技有限公司 | The method, apparatus and electronic equipment of information recommendation |
| CN109002488A (en) * | 2018-06-26 | 2018-12-14 | 北京邮电大学 | A kind of recommended models training method and device based on first path context |
| CN109241403A (en) * | 2018-08-03 | 2019-01-18 | 腾讯科技（深圳）有限公司 | Item recommendation method, device, machinery equipment and computer readable storage medium |
| CN109460513A (en) * | 2018-10-31 | 2019-03-12 | 北京字节跳动网络技术有限公司 | Method and apparatus for generating clicking rate prediction model |
| CN109658120A (en) * | 2017-10-12 | 2019-04-19 | 腾讯科技（深圳）有限公司 | A kind of business data processing method and device |
| CN109902753A (en) * | 2019-03-06 | 2019-06-18 | 深圳市珍爱捷云信息技术有限公司 | User's recommended models training method, device, computer equipment and storage medium |
-
2019
- 2019-07-01 CN CN201910584353.3A patent/CN110324418B/en active Active
Patent Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103150663A (en) * | 2013-02-18 | 2013-06-12 | 亿赞普(北京)科技有限公司 | Method and device for placing network placement data |
| CN103955545A (en) * | 2014-05-22 | 2014-07-30 | 成都品果科技有限公司 | Personalized social network influence identifying method |
| US20160241605A1 (en) * | 2015-02-17 | 2016-08-18 | Google Inc. | Second-screen control automatic pairing using push notifications |
| CN106570014A (en) * | 2015-10-09 | 2017-04-19 | 阿里巴巴集团控股有限公司 | Method and device for determining home attribute information of user |
| US20180150464A1 (en) * | 2016-11-30 | 2018-05-31 | Facebook, Inc. | Multi-stage Filtering for Recommended User Connections on Online Social Networks |
| CN107103057A (en) * | 2017-04-13 | 2017-08-29 | 腾讯科技（深圳）有限公司 | A kind of resource supplying method and device |
| CN107464141A (en) * | 2017-08-07 | 2017-12-12 | 北京京东尚科信息技术有限公司 | For the method, apparatus of information popularization, electronic equipment and computer-readable medium |
| CN107895277A (en) * | 2017-09-30 | 2018-04-10 | 平安科技（深圳）有限公司 | Method, electronic installation and the medium of push loan advertisement in the application |
| CN109658120A (en) * | 2017-10-12 | 2019-04-19 | 腾讯科技（深圳）有限公司 | A kind of business data processing method and device |
| CN108460590A (en) * | 2018-02-06 | 2018-08-28 | 北京三快在线科技有限公司 | The method, apparatus and electronic equipment of information recommendation |
| CN109002488A (en) * | 2018-06-26 | 2018-12-14 | 北京邮电大学 | A kind of recommended models training method and device based on first path context |
| CN109241403A (en) * | 2018-08-03 | 2019-01-18 | 腾讯科技（深圳）有限公司 | Item recommendation method, device, machinery equipment and computer readable storage medium |
| CN109460513A (en) * | 2018-10-31 | 2019-03-12 | 北京字节跳动网络技术有限公司 | Method and apparatus for generating clicking rate prediction model |
| CN109902753A (en) * | 2019-03-06 | 2019-06-18 | 深圳市珍爱捷云信息技术有限公司 | User's recommended models training method, device, computer equipment and storage medium |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111340605A (en) * | 2020-05-22 | 2020-06-26 | 支付宝(杭州)信息技术有限公司 | Method and device for training user behavior prediction model and user behavior prediction |
| CN112243021A (en) * | 2020-05-25 | 2021-01-19 | 北京沃东天骏信息技术有限公司 | Information pushing method, device, equipment and computer readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110324418B (en) | 2022-09-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Amin et al. | Customer churn prediction in telecommunication industry using data certainty | |
| Åstebro et al. | Predictors of the survival of innovations | |
| Zeinalizadeh et al. | Modeling and analysis of bank customer satisfaction using neural networks approach | |
| KR100751965B1 (en) | method and system for predicting attrition customers | |
| JP2004528628A (en) | Artificial intelligence trend system | |
| Cao et al. | Hybrid computational model for forecasting Taiwan construction cost index | |
| CN104321794A (en) | A system and method using multi-dimensional rating to determine an entity's future commercial viability | |
| CN112668859A (en) | Big data based customer risk rating method, device, equipment and storage medium | |
| CN106408325A (en) | User consumption behavior prediction analysis method based on user payment information and system | |
| CN102667775A (en) | Method for training and using a classification model with association rule models | |
| Chen et al. | A combined mining-based framework for predicting telecommunications customer payment behaviors | |
| Rachid et al. | Clustering prediction techniques in defining and predicting customers defection: The case of e-commerce context | |
| CN110324418A (en) | Method and apparatus based on customer relationship transmission service | |
| CN110310123A (en) | Risk judgment method and apparatus | |
| Guliyev et al. | Customer churn analysis in banking sector: Evidence from explainable machine learning model | |
| Hosseini et al. | Identifying multi-channel value co-creator groups in the banking industry | |
| Iranmanesh et al. | Customer churn prediction using artificial neural network: An analytical CRM application | |
| Giri et al. | Biogeography based optimization for mining rules to assess credit risk | |
| CN115545886A (en) | Overdue risk identification method, overdue risk identification device, overdue risk identification equipment and storage medium | |
| Hagen et al. | Processes, potential benefits, and limitations of big data analytics: a case analysis of 311 data from city of Miami | |
| Wagh et al. | Customer churn prediction in telecom sector using machine learning techniques | |
| Monfared et al. | Using business intelligence capabilities to improve the quality of decision-making: a case study of Mellat bank | |
| CN107871286A (en) | User is with contacting human world cohesion decision method/system, storage medium and equipment | |
| CN111199473A (en) | Anti-cheating method, device and system based on transaction record information | |
| CN111192045A (en) | Anti-cheating method, device and system based on transaction record information |
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 | ||
| REG | Reference to a national code |
Ref country code: HK Ref legal event code: DE Ref document number: 40015620 Country of ref document: HK |
|
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20200923 Address after: Cayman Enterprise Centre, 27 Hospital Road, George Town, Grand Cayman, British Islands Applicant after: Advanced innovation technology Co.,Ltd. Address before: A four-storey 847 mailbox in Grand Cayman Capital Building, British Cayman Islands Applicant before: Alibaba Group Holding Ltd. Effective date of registration: 20200923 Address after: Cayman Enterprise Centre, 27 Hospital Road, George Town, Grand Cayman, British Islands Applicant after: Innovative advanced technology Co.,Ltd. Address before: Cayman Enterprise Centre, 27 Hospital Road, George Town, Grand Cayman, British Islands Applicant before: Advanced innovation technology Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |