CN107507016A - A kind of information push method and system - Google Patents
A kind of information push method and system Download PDFInfo
- Publication number
- CN107507016A CN107507016A CN201710517520.3A CN201710517520A CN107507016A CN 107507016 A CN107507016 A CN 107507016A CN 201710517520 A CN201710517520 A CN 201710517520A CN 107507016 A CN107507016 A CN 107507016A
- Authority
- CN
- China
- Prior art keywords
- portrait
- fraction
- group
- classification
- classifications
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- 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
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/01—Social networking
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L12/00—Data switching networks
- H04L12/02—Details
- H04L12/16—Arrangements for providing special services to substations
- H04L12/18—Arrangements for providing special services to substations for broadcast or conference, e.g. multicast
- H04L12/1859—Arrangements for providing special services to substations for broadcast or conference, e.g. multicast adapted to provide push services, e.g. data channels
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
-
- 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
The embodiments of the invention provide a kind of information push method and system, wherein methods described includes：It is determined that at least one similar users group corresponding with targeted customer group；Based on the similar users group, multiple candidate's classifications are selected from specified category list；Based on the current contextual information of targeted customer group, the target classification for being recommended is determined from the multiple candidate's classification.The embodiment of the present invention can combine real-time environmental information and similar users group, the target classification for recommending more to match for targeted customer group, have higher accuracy, and less dependence manual intervention.
Description
Technical field
The present invention relates to technical field of data processing, more particularly to a kind of information push method and a kind of message push system
System.
Background technology
With the development of Internet information technique, O2O (Online To Offline, under line on online offline/line) business
Industry pattern is gradually risen.Wherein, O2O refers to be combined the commercial chance under line with internet, allows internet to turn into off-line transaction
Platform.
In O2O fields, user mainly has two kinds using the flatbed application program app approach for carrying out consumption decision：When
User is actively initiated the operation such as browse, screen, search for, and decision-making is carried out after acquisition relevant information；Second, platform by operation activity,
Form guiding user's decision-making of the message such as personalized recommendation push (PUSH).
Message push is the important means of app operations, can actively touch and reach user, current most commending system, main
To be directed to personal consumption feature to summarize, take out individual the emphasis demand and purchase intention, with reference to popular general
Time behavior trend COMPREHENSIVE CALCULATING go out the potential consumer motivation of user, so as to carry out correct guidance, reach the purpose precisely recommended.
However, with the popularization of social networks, personalized recommendation often have ignored the social demand of a colony, and group
Body Social behaviors are just gradually being evolved into the major part of individual consumption.The proposed algorithm of the existing comprehensive colony's Social behaviors of energy exists
Often there is certain limitation in machine learning and scene partitioning to more people's behaviors, make as the accuracy rate of intelligent recommendation is too low
Colony excessively intervenes recommendation condition, and situation elements are single can not to take polynary dimension etc. into account.
The content of the invention
In view of the above problems, it is proposed that the embodiment of the present invention overcomes above mentioned problem or at least in part to provide one kind
A kind of information push method and a kind of corresponding message push system to solve the above problems.
In order to solve the above problems, the embodiment of the invention discloses a kind of information push method, including：
It is determined that at least one similar users group corresponding with targeted customer group；
Based on the similar users group, multiple candidate's classifications are selected from specified category list；
Based on the current contextual information of targeted customer group, determined from the multiple candidate's classification for being recommended
Target classification.
Preferably, before the step of determination at least one similar users group corresponding with targeted customer group,
Also include：
Determine targeted customer group.
Preferably, methods described also includes：
Message push is carried out based on the target classification.
Preferably, the step of determination at least one similar users group corresponding with targeted customer group includes：
Obtain first portrait fraction of the targeted customer group for all classifications in the category list；
The coefficient of similarity of the first portrait fraction and multiple second portrait fractions in default portrait storehouse is calculated respectively, its
In, the multiple candidate user groups gathered in advance and each candidate user group are stored for described in the default portrait storehouse
Second portrait fraction of all classifications in category list；
Based on the coefficient of similarity, at least one similar users group is filtered out from the multiple candidate user group
Group.
Preferably, first portrait point for obtaining the targeted customer group and being directed to all classifications in the category list
Several steps include：
Obtain each historical behavior data of the member in preset time period in the targeted customer group；
Based on the historical behavior data of each member, it is all in the category list to determine that each member is directed to respectively
The portrait fraction of classification；
The portrait fraction that all classifications in the category list are directed to using each member builds the targeted customer group
For the first portrait fraction of all classifications in the category list.
Preferably, the historical behavior data comprise at least navigation patterns data and trading activity data；It is described to be based on
The historical behavior data of each member, determine that each member is directed to the portrait point of all classifications in the category list respectively
Several steps include：
Using single member in the targeted customer group as dimension, for each classification in the category list, from institute
The member in historical behavior extracting data preset time period is stated to go for the navigation patterns data of the classification and transaction
For data；
Calculate the pageview of the navigation patterns data and the trading volume of the trading activity data in preset time period；
The first weight and the second weight are configured to the pageview and the trading volume respectively, and carry out summation fortune
Calculate, obtain portrait fraction of the member to the classification；
The portrait fraction of the classification is standardized and normalized.
Preferably, first weight and second weight determine in the following way：
Collecting sample data；
Extract behavior characteristic information from the sample data, the characteristic information comprise at least navigation patterns information and
Trading activity information；
The navigation patterns information and the trading activity information are trained using default machine learning algorithm；
During training, when reaching the minimum value of default loss function, the navigation patterns information pair is obtained
Second weight corresponding to the first weight and the trading activity information answered.
Preferably, the portrait fraction that all classifications in the category list are directed to using each member builds the mesh
Mark groups of users in the category list the step of the first portrait fraction of all classifications for including：
Each member based on the targeted customer group is directed to the portrait fraction of all classifications in the category list, raw
Into candidate's portrait matrix；
The exceptional value in candidate's portrait matrix is determined, wherein, the exceptional value includes some member to some classification
Portrait fraction be less than default abnormal thresholding portrait fraction；
Exceptional value in candidate portrait matrix is replaced with into predetermined threshold value, the targeted customer group is obtained and is directed to institute
State the first portrait fraction of all classifications in category list.
Preferably, it is described to be based on the coefficient of similarity, filtered out from the multiple candidate user group at least one
The step of similar users group, includes：
The candidate user that coefficient of similarity is more than default similarity thresholding is filtered out from the multiple candidate user group
Group, as similar users group.
Preferably, it is described to be based on the similar users group, multiple candidate's classifications are selected from specified category list
Step includes：
The second portrait fraction and corresponding coefficient of similarity based on the similar users group, calculate the class respectively
The final portrait fraction of each classification in mesh list；
Based on the final portrait fraction, multiple classifications are chosen from all classifications as candidate's classification.
Preferably, the contextual information includes at least one of following information：
The geographical position of each member in the targeted customer group；
Real-time temporal information；
Real-time climatic information.
The embodiment of the invention also discloses a kind of message push system, including：
Similar group determination module, for determining at least one similar users group corresponding with targeted customer group；
Candidate's classification chooses module, for based on the similar users group, being selected from specified category list multiple
Candidate's classification；
Target classification determining module, for based on the current contextual information of targeted customer group, from the multiple candidate's class
The target classification for being recommended is determined in mesh.
Preferably, the system also includes：
Target group determining module, for determining targeted customer group.
Preferably, in addition to：Message pushing module, for carrying out message push based on the target classification.
Preferably, the similar group determination module includes：
First portrait fraction acquisition submodule, for obtaining the targeted customer group for owning in the category list
First portrait fraction of classification；
Coefficient of similarity calculating sub module, for calculating the first portrait fraction and in default portrait storehouse multiple the respectively
The coefficient of similarity of two portrait fractions, wherein, store the multiple candidate user groups gathered in advance in the default portrait storehouse
And each candidate user group is directed to the second portrait fraction of all classifications in the category list；
Submodule screens in similar group, for based on the coefficient of similarity, being sieved from the multiple candidate user group
Select at least one similar users group.
Preferably, the first portrait fraction acquisition submodule includes：
Historical data acquiring unit, for obtaining each member's going through in preset time period in the targeted customer group
History behavioral data；
Member's portrait determining unit, for the historical behavior data based on each member, each member is determined respectively
For the portrait fraction of all classifications in the category list；
Group's portrait construction unit, for being directed to the portrait fraction of all classifications in the category list using each member
Build first portrait fraction of the targeted customer group for all classifications in the category list.
Preferably, the historical behavior data comprise at least navigation patterns data and trading activity data；The member
Portrait determining unit includes：
Subelement is screened in behavior, for using single member in the targeted customer group as dimension, being arranged for the classification
Each classification in table, the member the browsing for the classification out of described historical behavior extracting data preset time period
Behavioral data and trading activity data；
Computation subunit, for calculating the pageview of the navigation patterns data and the transaction row in preset time period
For the trading volume of data；
Weight configures subelement, for configuring the first weight and second to the pageview and the trading volume respectively
Weight, and summation operation is carried out, obtain portrait fraction of the member to the classification；
Subelement is handled, for being standardized to the portrait fraction of the classification and normalized.
Preferably, first weight and second weight determine in the following way：
Collecting sample data；
Extract behavior characteristic information from the sample data, the characteristic information comprise at least navigation patterns information and
Trading activity information；
The navigation patterns information and the trading activity information are trained using default machine learning algorithm；
During training, when reaching the minimum value of default loss function, the navigation patterns information pair is obtained
Second weight corresponding to the first weight and the trading activity information answered.
Preferably, group's portrait construction unit includes：
Candidate's portrait matrix generation subelement, the classification is directed to for each member based on the targeted customer group
The portrait fraction of all classifications in list, generation candidate's portrait matrix；
Exceptional value determination subelement, for determining the exceptional value in candidate's portrait matrix, wherein, the exceptional value bag
Include the portrait fraction that some member is less than default abnormal thresholding to the portrait fraction of some classification；
Outlier processing subelement, predetermined threshold value is replaced with for the exceptional value in matrix that the candidate draws a portrait, is obtained
The targeted customer group is directed to the first portrait fraction of all classifications in the category list.
Preferably, similar group's screening submodule is additionally operable to：
The candidate user that coefficient of similarity is more than default similarity thresholding is filtered out from the multiple candidate user group
Group, as similar users group.
Preferably, candidate's classification is chosen module and included：
Final portrait calculating sub module, for the second portrait fraction based on the similar users group and corresponding phase
Like coefficient is spent, the final portrait fraction of each classification in the category list is calculated respectively；
Candidate's classification determination sub-module, for based on the final portrait fraction, multiple classifications to be chosen from all classifications
As candidate's classification.
Preferably, the contextual information includes at least one of following information：
The geographical position of each member in the targeted customer group；
Real-time temporal information；
Real-time climatic information.
The embodiment of the invention also discloses a kind of electronic equipment, including memory, processor and storage are on a memory simultaneously
The computer program that can be run on a processor, it is characterised in that realize the above method during computing device described program
The step of.
The embodiment of the invention also discloses a kind of computer-readable recording medium, computer program is stored thereon with, it is special
The step of sign is, the program realizes the above method when being executed by processor.
The embodiment of the present invention includes advantages below：
In embodiments of the present invention, real-time environmental information and similar users group can be combined, is potential user group
Group recommends the target classification more matched, has higher accuracy, and less dependence manual intervention.
The classification that the embodiment of the present invention is applied to special group is recommended, and recommends its for group behavior feature interested
Classification, the social attribute of colony is enhanced to a certain extent.
Brief description of the drawings
Fig. 1 is a kind of step flow chart of information push method embodiment one of the present invention；
Fig. 2 is a kind of step flow chart of information push method embodiment two of the present invention；
Fig. 3 is a kind of attenuation function curve synoptic diagram of information push method embodiment two of the present invention；
Fig. 4 is a kind of structured flowchart of message push system embodiment of the present invention.
Embodiment
In order to facilitate the understanding of the purposes, features and advantages of the present invention, it is below in conjunction with the accompanying drawings and specific real
Applying mode, the present invention is further detailed explanation.
One of the core concepts of the embodiments of the present invention is, by the historical behavior data of targeted customer group, builds mesh
The first portrait fraction of groups of users is marked, and similar multiple similar users groups are found based on collaborative filtering, is chosen
TopN classification is as candidate's classification set.The factor such as generalized time, place, weather chooses mesh from candidate's classification set simultaneously
The recommendation that classification carries out related trade company is marked, the social attribute of user community is enhanced well, lifts user experience.
Reference picture 1, a kind of step flow chart of information push method embodiment one of the present invention is shown, can specifically be wrapped
Include following steps：
Step 101, it is determined that at least one similar users group corresponding with targeted customer group；
Step 102, based on the similar users group, multiple candidate's classifications are selected from specified category list；
Step 103, based on the current contextual information of targeted customer group, determine to be used for from the multiple candidate's classification into
The target classification that row is recommended.
In embodiments of the present invention, when it is determined that after at least one similar users group corresponding with targeted customer group,
Candidate's classification, and the contextual information that combining target groups of users is current can be determined based on the similar users group, from multiple times
Select the target classification determined in classification for being recommended, classification push on, it is contemplated that similar users group and in real time
Contextual information, improve the accuracy rate of recommendation.
Reference picture 2, a kind of step flow chart of information push method embodiment two of the present invention is shown, can specifically be wrapped
Include following steps：
Step 201, targeted customer group is determined；
In the specific implementation, the embodiment of the present invention can apply in specified application, the specified application can be with
Possess the function of establishing group, after user logs in the application program, group can be established by establishing the function of group, and
Multiple good friend users are pulled in into group, obtain targeted customer group.
Targeted customer group is the targeted user population for needing to carry out message recommendation.As a kind of example, potential user group
Group can include but is not limited to the information such as the login account of each member, the pet name, positional information.
Step 202, it is determined that at least one similar users group corresponding with the targeted customer group；
After targeted customer group is obtained, the embodiment of the present invention may further determine that corresponding to the targeted customer group
One or more similar users groups.
In a kind of preferred embodiment of the embodiment of the present invention, step 202 can further include following sub-step：
Sub-step S11, obtain first portrait point of the targeted customer group for all classifications in the category list
Number；
Specifically, portrait be according to information such as the social property of user, habits and customs and consumer behaviors and take out one
The user model of individual labeling.In embodiments of the present invention, the first portrait fraction can refer to what is obtained according to above-mentioned user model
Targeted customer group specified category list is analyzed after numerical value.
In practice, a variety of classifications can be included in the category list specified, a variety of classifications can be one-level classification,
Can be two level classification, for example, the category list can include all two level classification under cuisines category, such as western-style food, Chinese meal,
Chafing dish, barbecue, cooking, buffet etc..
In a kind of preferred embodiment of the embodiment of the present invention, sub-step S11 can further include following sub-step：
Sub-step S111, obtain each historical behavior number of the member in preset time period in the targeted customer group
According to；
In the specific implementation, each member can be obtained out of targeted customer group in log recording corresponding to each member
Historical behavior data.
As a kind of example, the historical behavior data can include but is not limited to navigation patterns data (page view, letter
Claim pv), click on behavioral data, trading activity data etc..
It should be noted that preset time period can be set according to the actual requirements, for example, can set preset time period
For nearest 30 days, the embodiment of the present invention was not construed as limiting to this.
Sub-step S112, based on the historical behavior data of each member, determine that each member is directed to the class respectively
The portrait fraction of all classifications in mesh list；
, can be to the historical behavior data when collecting in targeted customer group after the historical behavior data of each member
Analyzed, to determine that each member is directed to the portrait fraction of all classifications in the category list, i.e., each member is to specifying
Category list in each classification numerical value for obtained after behavioural analysis collect.
In a kind of preferred embodiment of the embodiment of the present invention, sub-step S112 can further include following sub-step：
Sub-step S1121, using single member in the targeted customer group as dimension, for every in the category list
Individual classification, the member is directed to the navigation patterns data of the classification out of described historical behavior extracting data preset time period
And trading activity data；
Specifically, when collecting in targeted customer group after the historical behavior data of each member, can with it is single into
Member is dimension, and for each classification, navigation patterns data of the member to each classification are filtered out from the historical behavior data
And trading activity data.
Sub-step S1122, calculate the pageview of the navigation patterns data and the trading activity in preset time period
The trading volume of data；
Sub-step S1121 determine single member to the navigation patterns data of each classification and trading activity data with
Afterwards, further the navigation patterns data and trading activity data can be counted, obtains the member in the preset time
The pageview and trading volume of each classification are directed in section.
In the implementation, pageview can use equation below to calculate：
Wherein, piThe pv amounts apart from i days today are represented, I can be the numerical value of preset time period, for example, I=30.
Trading volume can use equation below to calculate：
Wherein, OiThe trading volume apart from i days today is represented, I can be the numerical value of preset time period, for example, I=30.
Sub-step S1123, the first weight and the second weight are configured to the pageview and the trading volume respectively, and
Summation operation is carried out, obtains portrait fraction of the member to the classification；
After single member is obtained to the pageview and trading volume of each classification, the first power can be configured to pageview
Weight, and, the second weight is configured for trading volume, then, summation operation is carried out to the pageview after configuration weight and trading volume, obtained
To the member to such purpose portrait fraction.
In one embodiment, the first weight and the second weight can determine in the following way：
Collecting sample data；Behavior characteristic information is extracted from the sample data, the characteristic information comprises at least clear
Look at behavioural information and trading activity information；Using default machine learning algorithm to the navigation patterns information and the friendship
Easy behavioural information is trained；During training, when reaching the minimum value of default loss function, described browse is obtained
Second weight corresponding to first weight corresponding to behavioural information and the trading activity information.
Specifically, log recording in a period of specified application can be gathered as sample data.
After collecting sample data, behavior characteristic information can be extracted from the sample data.
As a kind of example, this feature information can at least include navigation patterns information and trading activity information etc..
In the specific implementation, the sample data can include positive sample and negative sample, if from sample chooses day
Access times of the user based on some classification are more than the access thresholds for classification setting in following a period of time, then it is assumed that the sample
This is positive sample, whereas if access times are less than or equal to access thresholds, then it is assumed that the sample is negative sample.
According to the behavior characteristic information of user, the initial experience of navigation patterns information and trading activity information is set respectively
Value coefficient (initial weight)；For example, the initial experience value coefficient for setting navigation patterns information sets trading activity information as 0.8
Initial experience value coefficient be 0.2.
Setting time attenuation function is f (x)=e-ax；
Wherein, a is time attenuation coefficient, and x represents that within a period of time of sample selection day daily Distance Time section is newest
The date on date is poor, whenWhen function curve as shown in the schematic diagram of fig. 3, from figure 3, it can be seen that when x it is smaller, influence to imitate
Fruit is bigger.
After collecting sample, LR (Logistic Regression, logistic regression) algorithm can be based on and carry out engineering
Practise, obtain the first weight and the second weight.
Wherein, LR algorithm can be expressed as：
Wherein,X1,...,XnRepresent each behavioural characteristic of some sample data
Information.
In the training process of weight model, can set loss function as
Wherein, yiThe actual value of sample is represented, i.e., under truth, the sample data is positive sample data or negative sample
Data；hθ(x) predicted value of sample is represented, i.e., it is positive sample data or negative sample data to obtain the sample according to LR algorithm.
According to Algorithm Learning so that loss function is minimum value, finally give the first weight corresponding to navigation patterns information with
And the second weight corresponding to trading activity information.
For example, it is 0.3 that the first weight finally determined, which is the 0.7, second weight,.
It should be noted that the embodiment of the present invention is not limited to the mode of the above-mentioned weight of determination first and the second weight,
Those skilled in the art are set or calculated using other modes according to the actual requirements the first weight and the second weight and are possible.
After determining the first weight and the second weight, in one embodiment, can using it is linear plus and by the way of
Determine portrait fraction of the user to some classification.
For example, when it is 0.3 that the first weight, which is the 0.7, second weight, equation below can be used to calculate user to some class
Purpose portrait fraction：
Sub-step S1124, is standardized and normalized to the portrait fraction of the classification.
After user is obtained to the portrait fraction of some classification, the fraction can also be standardized and be normalized
Processing.
Specifically, adding and in the presence of extreme-value problem to a certain extent due to linear, therefore place can be standardized
Reason is by data bi-directional scaling, to be allowed to fall into a less specific section.
In one embodiment, can be standardized using equation below：
Wherein, x represent the linear of above-mentioned each classification plus and result,For the portrait of all classifications based on user's dimension
The average value of fraction, σXIt is then the standard deviation of the portrait fraction of all classifications based on user's dimension.
It the result obtained after standardization, can further be normalized, it is fallen in the range of [0,10].
In one embodiment, can be normalized using sigmoid functions, sigmoid function representations are such as
Under：
The final result obtained by above-mentioned processing procedure, final score that can be as user to each classification.
After user is obtained to the portrait fraction of each classification, portrait fraction tissue that can be by user to all classifications
Get up, obtain portrait fraction of the user for all classifications in the category list.
For example, classification includes " buffet ", " Japanese cuisine ", three kinds of " western-style food ", portrait point of the user to these three classifications
Number is respectively 2,4,7.Then the portrait fraction of the user is (2,4,7).
Sub-step S113, the portrait fraction that all classifications in the category list are directed to using each member build the mesh
Mark the first portrait fraction that groups of users is directed to all classifications in the category list.
Specifically, after obtaining each member for the portrait fraction of all classifications in the category list, can build
The two-dimensional matrix of " classification-member ".
In a kind of preferred embodiment of the embodiment of the present invention, sub-step S113 can further include following sub-step：
Sub-step S1131, each member based on the targeted customer group are directed to all classifications in the category list
Portrait fraction, generation candidate draw a portrait matrix；
It can represent as follows in the matrix in the specific implementation, candidate draws a portrait：
Wherein,Represent that user c is in the portrait fraction of k classifications in targeted customer group i.
Sub-step S1132, determine the exceptional value that the candidate draws a portrait in matrix, wherein, the exceptional value include some into
Member is less than the portrait fraction of default abnormal thresholding to the portrait fraction of some classification；
In the specific implementation, for all classifications, some member is there may be to some classification in targeted customer group
The relatively low situation of preference, now, the member is drawn a portrait to such purpose, and generally than relatively low, (the usual portrait fraction can be far below fraction
The mean value level of candidate's portrait matrix), in embodiments of the present invention, the member can be worked as to the portrait fraction of this classification
Handled as exceptional value, to cause the first final portrait fraction to meet the hobby need of most people in targeted customer group
Ask.
In one embodiment, exceptional value can include some member to the portrait fraction of some classification less than default
The portrait fraction of abnormal thresholding.
As a kind of example, default abnormal thresholding can use equation below to calculate：
Wherein,For in targeted customer group all classifications of c user portrait fraction average value,Used for target
The standard deviation of the portrait fraction of all classifications of c user in the group of family.
Sub-step S1133, the exceptional value in candidate portrait matrix is replaced with into predetermined threshold value, the target is obtained and uses
Family group is directed to the first portrait fraction of all classifications in the category list.
After determining the exceptional value in candidate's portrait matrix, the exceptional value can be replaced with to predetermined threshold value, and will be abnormal
The portrait matrix that value obtains after replacing is directed to the first portrait point of all classifications in the category list as targeted customer group
Number.
In one embodiment, the predetermined threshold value can be set as numerical value 0, then corresponding abnormality processing algorithm can be with
For：
Sub-step S12, the phase of the first portrait fraction and multiple second portrait fractions in default portrait storehouse is calculated respectively
Like degree coefficient；
In embodiments of the present invention, the groups of users established can also be gathered in advance as candidate user group, and is adopted
Collect the second portrait fraction of the candidate user group, establish portrait storehouse, be i.e. the portrait storehouse can include the multiple times gathered in advance
Groups of users and each candidate user group is selected to be directed to the second portrait fraction of all classifications in the category list.
Wherein, the second portrait fraction can be each member in candidate user group respectively to institute in specified category list
There is the portrait fraction of classification, the determination mode of the second portrait fraction is identical with the determination mode of the first portrait fraction, may be referred to
The acquisition methods of above-mentioned first portrait fraction, will not be repeated here.
After obtaining candidate user group and its corresponding second portrait fraction in portrait storehouse, the embodiment of the present invention can
To calculate the coefficient of similarity of the first portrait fraction and each second portrait fraction in storehouse of drawing a portrait respectively.
In one embodiment, can use collaborative filtering calculate the first portrait fraction with each the in portrait storehouse
The coefficient of similarity of two portrait fractions, one of which collaborative filtering example are as follows：
Wherein, the second portrait point of sim (i, j) is targeted customer group i the first portrait fraction and candidate user group j
Several coefficient of similarity, value is bigger, represents that its similarity is higher.pikRepresent the portrait fraction of classification k in targeted customer group i
(i.e. summation of all users to classification k portrait fraction in targeted customer group i),For all classes in targeted customer group i
The average value of purpose portrait fraction；pjkThe portrait fraction of classification k in candidate user group j is represented,For in candidate user group j
The average value of the portrait fraction of all classifications.
Sub-step S13, based on the coefficient of similarity, one or more is filtered out from the multiple candidate user group
Similar users group.
Obtain in the first portrait fraction and portrait storehouse after the coefficient of similarity of multiple second portraits fractions, can be according to this
Coefficient of similarity, candidate user group similar with targeted customer group in portrait storehouse is found, that is, finds optimal neighborhood.
In one embodiment, can by coefficient of similarity will draw a portrait storehouse in it is all second portrait fractions be ranked up,
And choose sequence preceding (coefficient of similarity sorts from big to small) or after being ordered into (coefficient of similarity sorts from small to large) the
Candidate user group corresponding to two portrait fractions, as similar users group.
In another embodiment, similarity thresholding can be set, and coefficient of similarity is more than similarity thresholding
Candidate user group corresponding to second portrait fraction, as similar users group.
Step 203, based on the similar users group, multiple candidate's classifications are selected from specified category list；
After determining similar users group, can be based on similar users groups second draw a portrait fraction, specified from all
Multiple candidate's classifications are chosen in category list.
In a kind of preferred embodiment of the embodiment of the present invention, step 203 can include following sub-step：
Sub-step S21, the second portrait fraction and corresponding similar based on one or more of similar users groups
Coefficient is spent, calculates the final portrait fraction of each classification in the category list respectively；
In one embodiment, can be according to corresponding to the similar users group after determining similar users group
Two portrait fractions and coefficient of similarity, the final portrait fraction of each classification is calculated using following algorithm：
Wherein, pjkFor the portrait fraction of k classifications in similar users group j, sim (i, j) is the first of targeted customer group i
The coefficient of similarity drawn a portrait between fraction and similar users group j the second portrait fraction.
Sub-step S22, based on the final portrait fraction, multiple classifications are chosen from all classifications as candidate's classification.
In one embodiment, obtain in category list after the final portrait fraction of all classifications, can be by this most
Portrait fraction is ranked up in descending order or ascending order eventually, and is chosen sequence at preceding (descending) or sorted in multiple classes of rear (ascending order)
Mesh, as candidate's classification.
In another embodiment, score threshold can be set, final portrait fraction is more than to the class of the score threshold
Mesh is as candidate's classification.
Step 204, based on the current contextual information of targeted customer group, determine to be used for from the multiple candidate's classification into
The target classification that row is recommended.
In practice, candidate's classification can be multiple.In embodiments of the present invention, when it is determined that multiple candidate's classifications with
Afterwards, context aware factor can be added, personalized classification is completed to targeted customer group according to current special scenes and recommended.
The context aware factor can include the contextual information of the current environment obtained in real time.
As a kind of example of the embodiment of the present invention, the contextual information can include at least one of following information：
(1) in the targeted customer group each member geographical position；
In the specific implementation, the geographical position of each member in targeted customer group by positioning function, can be obtained.
(2) real-time temporal information；
In the specific implementation, the temporal information can determine that current time is early, middle and late and working day, festivals or holidays etc.
Time type.
(3) real-time climatic information.
Specifically, climatic information can include temperature information, weather information etc..
, can be according to the contextual information after obtaining real-time contextual information, the selection target class from multiple candidate's classifications
Mesh, and target classification is generated into PUSH message by specified application and is pushed in targeted customer group.
In practice, after obtaining target classification, the embodiment of the present invention can determine that trade company corresponding to the target classification enters
Row message pushes.
For example, LBS (Location Based can be based on by the geographical position of each member in targeted customer group
Service, location Based service) service progress commercial circle recommendation；And for example, by the morning, afternoon and evening and the time such as festivals or holidays on working day
Factor recommends different types of trade company；Or trade company is recommended according to factors such as current season, weather.
In embodiments of the present invention, real-time contextual information and similar users group can be combined, is potential user group
Group recommends the target classification more matched, has higher accuracy, and less dependence manual intervention.
The classification that the embodiment of the present invention is applied to special group is recommended, and recommends its for group behavior feature interested
Classification, the social attribute of colony is enhanced to a certain extent.
It should be noted that for embodiment of the method, in order to be briefly described, therefore it is all expressed as to a series of action group
Close, but those skilled in the art should know, the embodiment of the present invention is not limited by described sequence of movement, because according to
According to the embodiment of the present invention, some steps can use other orders or carry out simultaneously.Secondly, those skilled in the art also should
Know, embodiment described in this description belongs to preferred embodiment, and the involved action not necessarily present invention is implemented
Necessary to example.
Reference picture 4, a kind of structured flowchart of message push system embodiment of the present invention is shown, can specifically included such as
Lower module：
Similar group determination module 401, for determining at least one similar users group corresponding with targeted customer group；
Candidate's classification chooses module 402, for based on the similar users group, being selected from specified category list more
Individual candidate's classification；
Target classification determining module 403, for based on the current contextual information of targeted customer group, from the multiple candidate
Determine to be used for the target classification recommended in classification.
In a kind of preferred embodiment of the embodiment of the present invention, the system can also include following module：
Target group determining module, for determining targeted customer group.
In a kind of preferred embodiment of the embodiment of the present invention, the system can also include following module：
Message pushing module, for carrying out message push based on the target classification.
In a kind of preferred embodiment of the embodiment of the present invention, the similar group determination module 302 can include as follows
Submodule：
First portrait fraction acquisition submodule, for obtaining the targeted customer group for owning in the category list
First portrait fraction of classification；
Coefficient of similarity calculating sub module, for calculating the first portrait fraction and in default portrait storehouse multiple the respectively
The coefficient of similarity of two portrait fractions, wherein, store the multiple candidate user groups gathered in advance in the default portrait storehouse
And each candidate user group is directed to the second portrait fraction of all classifications in the category list；
Submodule screens in similar group, for based on the coefficient of similarity, being sieved from the multiple candidate user group
Select at least one similar users group.
In a kind of preferred embodiment of the embodiment of the present invention, the first portrait fraction acquisition submodule can be included such as
Lower unit：
Historical data acquiring unit, for obtaining each member's going through in preset time period in the targeted customer group
History behavioral data；
Member's portrait determining unit, for the historical behavior data based on each member, each member is determined respectively
For the portrait fraction of all classifications in the category list；
Group's portrait construction unit, for being directed to the portrait fraction of all classifications in the category list using each member
Build first portrait fraction of the targeted customer group for all classifications in the category list.
In a kind of preferred embodiment of the embodiment of the present invention, the historical behavior data comprise at least navigation patterns data
And trading activity data；Member's portrait determining unit can include following subelement：
Subelement is screened in behavior, for using single member in the targeted customer group as dimension, being arranged for the classification
Each classification in table, the member the browsing for the classification out of described historical behavior extracting data preset time period
Behavioral data and trading activity data；
Computation subunit, for calculating the pageview of the navigation patterns data and the transaction row in preset time period
For the trading volume of data；
Weight configures subelement, for configuring the first weight and second to the pageview and the trading volume respectively
Weight, and summation operation is carried out, obtain portrait fraction of the member to the classification；
Subelement is handled, for being standardized to the portrait fraction of the classification and normalized.
In a kind of preferred embodiment of the embodiment of the present invention, first weight and second weight are using as follows
Mode determines：
Collecting sample data；
Extract behavior characteristic information from the sample data, the characteristic information comprise at least navigation patterns information and
Trading activity information；
The navigation patterns information and the trading activity information are trained using default machine learning algorithm；
During training, when reaching the minimum value of default loss function, the navigation patterns information pair is obtained
Second weight corresponding to the first weight and the trading activity information answered.
In a kind of preferred embodiment of the embodiment of the present invention, it is single that group's portrait construction unit can include following son
Member：
Candidate's portrait matrix generation subelement, the classification is directed to for each member based on the targeted customer group
The portrait fraction of all classifications in list, generation candidate's portrait matrix；
Exceptional value determination subelement, for determining the exceptional value in candidate's portrait matrix, wherein, the exceptional value bag
Include the portrait fraction that some member is less than default abnormal thresholding to the portrait fraction of some classification；
Outlier processing subelement, predetermined threshold value is replaced with for the exceptional value in matrix that the candidate draws a portrait, is obtained
The targeted customer group is directed to the first portrait fraction of all classifications in the category list.
In a kind of preferred embodiment of the embodiment of the present invention, similar group's screening submodule is additionally operable to：
The candidate user that coefficient of similarity is more than default similarity thresholding is filtered out from the multiple candidate user group
Group, as similar users group.
In a kind of preferred embodiment of the embodiment of the present invention, candidate's classification selection module 303 can include as follows
Submodule：
Final portrait calculating sub module, for the second portrait fraction based on the similar users group and corresponding phase
Like coefficient is spent, the final portrait fraction of each classification in the category list is calculated respectively；
Candidate's classification determination sub-module, for based on the final portrait fraction, multiple classifications to be chosen from all classifications
As candidate's classification.
In a kind of preferred embodiment of the embodiment of the present invention, the contextual information includes at least one of following information：
The geographical position of each member in the targeted customer group；
Real-time temporal information；
Real-time climatic information.
For Fig. 3 system embodiment, because it is substantially similar to Fig. 1 embodiment of the method, so the ratio of description
Relatively simple, related part illustrates referring to the part of Fig. 1 embodiment of the method.
The embodiment of the invention also discloses a kind of electronic equipment, including memory, processor and storage are on a memory simultaneously
The computer program that can be run on a processor, Fig. 1 and/or Fig. 2 methods describeds are realized during the computing device described program
The step of.
The embodiment of the invention also discloses a kind of computer-readable recording medium, computer program is stored thereon with, the journey
The step of Fig. 1 and/or Fig. 2 methods describeds are realized when sequence is executed by processor.
Each embodiment in this specification is described by the way of progressive, what each embodiment stressed be with
The difference of other embodiment, between each embodiment identical similar part mutually referring to.
It should be understood by those skilled in the art that, the embodiment of the embodiment of the present invention can be provided as method, apparatus or calculate
Machine program product.Therefore, the embodiment of the present invention can use complete hardware embodiment, complete software embodiment or combine software and
The form of the embodiment of hardware aspect.Moreover, the embodiment of the present invention can use one or more wherein include computer can
With in the computer-usable storage medium (including but is not limited to magnetic disk storage, CD-ROM, optical memory etc.) of program code
The form of the computer program product of implementation.
The embodiment of the present invention is with reference to method according to embodiments of the present invention, terminal device (system) and computer program
The flow chart and/or block diagram of product describes.It should be understood that can be by computer program instructions implementation process figure and/or block diagram
In each flow and/or square frame and the flow in flow chart and/or block diagram and/or the combination of square frame.These can be provided
Computer program instructions are set to all-purpose computer, special-purpose computer, Embedded Processor or other programmable data processing terminals
Standby processor is to produce a machine so that is held by the processor of computer or other programmable data processing terminal equipments
Capable instruction is produced for realizing in one flow of flow chart or multiple flows and/or one square frame of block diagram or multiple square frames
The device for the function of specifying.
These computer program instructions, which may be alternatively stored in, can guide computer or other programmable data processing terminal equipments
In the computer-readable memory to work in a specific way so that the instruction being stored in the computer-readable memory produces bag
The manufacture of command device is included, the command device is realized in one flow of flow chart or multiple flows and/or one side of block diagram
The function of being specified in frame or multiple square frames.
These computer program instructions can be also loaded into computer or other programmable data processing terminal equipments so that
Series of operation steps is performed on computer or other programmable terminal equipments to produce computer implemented processing, so that
The instruction performed on computer or other programmable terminal equipments is provided for realizing in one flow of flow chart or multiple flows
And/or specified in one square frame of block diagram or multiple square frames function the step of.
Although having been described for the preferred embodiment of the embodiment of the present invention, those skilled in the art once know base
This creative concept, then other change and modification can be made to these embodiments.So appended claims are intended to be construed to
Including preferred embodiment and fall into having altered and changing for range of embodiment of the invention.
Finally, it is to be noted that, herein, such as first and second or the like relational terms be used merely to by
One entity or operation make a distinction with another entity or operation, and not necessarily require or imply these entities or operation
Between any this actual relation or order be present.Moreover, term " comprising ", "comprising" or its any other variant meaning
Covering including for nonexcludability, so that process, method, article or terminal device including a series of elements are not only wrapped
Those key elements, but also the other element including being not expressly set out are included, or is also included for this process, method, article
Or the key element that terminal device is intrinsic.In the absence of more restrictions, wanted by what sentence "including a ..." limited
Element, it is not excluded that other identical element in the process including the key element, method, article or terminal device also be present.
Above to a kind of information push method provided by the present invention and a kind of message push system, detailed Jie has been carried out
Continue, specific case used herein is set forth to the principle and embodiment of the present invention, and the explanation of above example is only
It is the method and its core concept for being used to help understand the present invention；Meanwhile for those of ordinary skill in the art, according to this hair
Bright thought, there will be changes in specific embodiments and applications, in summary, this specification content should not manage
Solve as limitation of the present invention.
Claims (24)
- A kind of 1. information push method, it is characterised in that including：It is determined that at least one similar users group corresponding with targeted customer group；Based on the similar users group, multiple candidate's classifications are selected from specified category list；Based on the current contextual information of targeted customer group, the target for being recommended is determined from the multiple candidate's classification Classification.
- 2. according to the method for claim 1, it is characterised in that determine corresponding with targeted customer group at least one described Before the step of individual similar users group, in addition to：Determine targeted customer group.
- 3. method according to claim 1 or 2, it is characterised in that also include：Message push is carried out based on the target classification.
- 4. according to the method for claim 1, it is characterised in that the determination is corresponding with targeted customer group at least one The step of similar users group, includes：Obtain first portrait fraction of the targeted customer group for all classifications in the category list；The coefficient of similarity of the first portrait fraction and multiple second portrait fractions in default portrait storehouse is calculated respectively, wherein, The multiple candidate user groups gathered in advance are stored in the default portrait storehouse and each candidate user group is directed to the class Second portrait fraction of all classifications in mesh list；Based on the coefficient of similarity, at least one similar users group is filtered out from the multiple candidate user group.
- 5. according to the method for claim 4, it is characterised in that the acquisition targeted customer group is directed to the classification Include in list the step of the first portrait fraction of all classifications：Obtain each historical behavior data of the member in preset time period in the targeted customer group；Based on the historical behavior data of each member, determine that each member is directed to all classifications in the category list respectively Portrait fraction；The portrait fraction structure targeted customer group that all classifications in the category list are directed to using each member is directed to First portrait fraction of all classifications in the category list.
- 6. according to the method for claim 5, it is characterised in that the historical behavior data comprise at least navigation patterns data And trading activity data；The historical behavior data based on each member, determine each member for described respectively Include in category list the step of the portrait fraction of all classifications：Using single member in the targeted customer group as dimension, for each classification in the category list, gone through from described Navigation patterns data and trading activity number that the member in preset time period is directed to the classification are extracted in history behavioral data According to；Calculate the pageview of the navigation patterns data and the trading volume of the trading activity data in preset time period；The first weight and the second weight are configured to the pageview and the trading volume respectively, and carry out summation operation, is obtained Obtain portrait fraction of the member to the classification；The portrait fraction of the classification is standardized and normalized.
- 7. according to the method for claim 6, it is characterised in that first weight and second weight are using as follows Mode determines：Collecting sample data；Behavior characteristic information is extracted from the sample data, the characteristic information comprises at least navigation patterns information and transaction Behavioural information；The navigation patterns information and the trading activity information are trained using default machine learning algorithm；During training, when reaching the minimum value of default loss function, obtain corresponding to the navigation patterns information Second weight corresponding to first weight and the trading activity information.
- 8. according to the method described in claim 5 or 6 or 7, it is characterised in that described to be arranged using each member for the classification The portrait fraction of all classifications builds the first picture that the targeted customer group is directed to all classifications in the category list in table As the step of fraction includes：Each member based on the targeted customer group is directed to the portrait fraction of all classifications in the category list, and generation is waited Choosing portrait matrix；The exceptional value in candidate's portrait matrix is determined, wherein, the exceptional value includes picture of some member to some classification As fraction is less than the portrait fraction of default abnormal thresholding；Exceptional value in candidate portrait matrix is replaced with into predetermined threshold value, the targeted customer group is obtained and is directed to the class First portrait fraction of all classifications in mesh list.
- 9. according to the method for claim 4, it is characterised in that it is described to be based on the coefficient of similarity, from the multiple time The step of filtering out at least one similar users group in groups of users is selected to include：The candidate user group that coefficient of similarity is more than default similarity thresholding is filtered out from the multiple candidate user group, As similar users group.
- 10. according to the method for claim 4, it is characterised in that it is described to be based on the similar users group, from specified class The step of multiple candidate's classifications are selected in mesh list includes：The second portrait fraction and corresponding coefficient of similarity based on the similar users group, calculate the classification row respectively The final portrait fraction of each classification in table；Based on the final portrait fraction, multiple classifications are chosen from all classifications as candidate's classification.
- 11. according to the method for claim 1, it is characterised in that the contextual information includes at least one of following information：The geographical position of each member in the targeted customer group；Real-time temporal information；Real-time climatic information.
- A kind of 12. message push system, it is characterised in that including：Similar group determination module, for determining at least one similar users group corresponding with targeted customer group；Candidate's classification chooses module, for based on the similar users group, multiple candidates to be selected from specified category list Classification；Target classification determining module, for based on the current contextual information of targeted customer group, from the multiple candidate's classification It is determined that for the target classification recommended.
- 13. system according to claim 12, it is characterised in that also include：Target group determining module, for determining targeted customer group.
- 14. the system according to claim 12 or 13, it is characterised in that also include：Message pushing module, for based on institute State target classification and carry out message push.
- 15. system according to claim 12, it is characterised in that the similar group determination module includes：First portrait fraction acquisition submodule, for obtaining the targeted customer group for all classifications in the category list First portrait fraction；Coefficient of similarity calculating sub module, drawn for calculating the first portrait fraction respectively with default portrait storehouse multiple second As the coefficient of similarity of fraction, wherein, multiple candidate user groups for gathering in advance and every are stored in the default portrait storehouse Individual candidate user group is directed to the second portrait fraction of all classifications in the category list；Submodule screens in similar group, for based on the coefficient of similarity, being filtered out from the multiple candidate user group At least one similar users group.
- 16. system according to claim 15, it is characterised in that the first portrait fraction acquisition submodule includes：Historical data acquiring unit, for obtaining each history row of the member in preset time period in the targeted customer group For data；Member's portrait determining unit, for the historical behavior data based on each member, determine that each member is directed to respectively The portrait fraction of all classifications in the category list；Group's portrait construction unit, for being directed to the portrait fraction structure of all classifications in the category list using each member The targeted customer group is directed to the first portrait fraction of all classifications in the category list.
- 17. system according to claim 16, it is characterised in that the historical behavior data comprise at least navigation patterns number According to this and trading activity data；Member's portrait determining unit includes：Subelement is screened in behavior, for using single member in the targeted customer group as dimension, in the category list Each classification, out of described historical behavior extracting data preset time period the member be directed to the classification navigation patterns Data and trading activity data；Computation subunit, for calculating the pageview of the navigation patterns data and the trading activity number in preset time period According to trading volume；Weight configures subelement, for configuring the first weight and the second power to the pageview and the trading volume respectively Weight, and summation operation is carried out, obtain portrait fraction of the member to the classification；Subelement is handled, for being standardized to the portrait fraction of the classification and normalized.
- 18. system according to claim 17, it is characterised in that first weight and second weight are using such as Under type determines：Collecting sample data；Behavior characteristic information is extracted from the sample data, the characteristic information comprises at least navigation patterns information and transaction Behavioural information；The navigation patterns information and the trading activity information are trained using default machine learning algorithm；During training, when reaching the minimum value of default loss function, obtain corresponding to the navigation patterns information Second weight corresponding to first weight and the trading activity information.
- 19. according to the system described in claim 16 or 17 or 18, it is characterised in that group's portrait construction unit includes：Candidate's portrait matrix generation subelement, the category list is directed to for each member based on the targeted customer group In all classifications portrait fraction, generation candidate draw a portrait matrix；Exceptional value determination subelement, for determining the exceptional value in candidate's portrait matrix, wherein, the exceptional value includes certain Individual member is less than the portrait fraction of default abnormal thresholding to the portrait fraction of some classification；Outlier processing subelement, predetermined threshold value is replaced with for the exceptional value in matrix that the candidate draws a portrait, is obtained described Targeted customer group is directed to the first portrait fraction of all classifications in the category list.
- 20. system according to claim 15, it is characterised in that similar group's screening submodule is additionally operable to：The candidate user group that coefficient of similarity is more than default similarity thresholding is filtered out from the multiple candidate user group, As similar users group.
- 21. system according to claim 15, it is characterised in that candidate's classification, which chooses module, to be included：Final portrait calculating sub module, for the second portrait fraction based on the similar users group and corresponding similarity Coefficient, the final portrait fraction of each classification in the category list is calculated respectively；Candidate's classification determination sub-module, for based on the final portrait fraction, multiple classification conducts to be chosen from all classifications Candidate's classification.
- 22. system according to claim 12, it is characterised in that the contextual information includes at least the one of following information Kind：The geographical position of each member in the targeted customer group；Real-time temporal information；Real-time climatic information.
- 23. a kind of electronic equipment, including memory, processor and storage are on a memory and the calculating that can run on a processor Machine program, it is characterised in that the step of any one of claim 1 to 11 methods described is realized during the computing device described program Suddenly.
- 24. a kind of computer-readable recording medium, is stored thereon with computer program, it is characterised in that the program is by processor The step of any one of claim 1 to 11 methods described is realized during execution.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710517520.3A CN107507016A (en) | 2017-06-29 | 2017-06-29 | A kind of information push method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710517520.3A CN107507016A (en) | 2017-06-29 | 2017-06-29 | A kind of information push method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107507016A true CN107507016A (en) | 2017-12-22 |
Family
ID=60678904
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710517520.3A Pending CN107507016A (en) | 2017-06-29 | 2017-06-29 | A kind of information push method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107507016A (en) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108537567A (en) * | 2018-03-06 | 2018-09-14 | 阿里巴巴集团控股有限公司 | A kind of determination method and apparatus of targeted user population |
| CN108734348A (en) * | 2018-05-14 | 2018-11-02 | 广东心里程教育集团有限公司 | A kind of method and system of automatic push online course |
| CN108898447A (en) * | 2018-07-23 | 2018-11-27 | 北京大前科技有限责任公司 | Obtain the method and device of extension content |
| CN109118282A (en) * | 2018-08-08 | 2019-01-01 | 福建百悦信息科技有限公司 | A kind of bimodulus mutual inductance intelligent space user draws a portrait management method and terminal |
| CN109389157A (en) * | 2018-09-14 | 2019-02-26 | 阿里巴巴集团控股有限公司 | A kind of user group recognition methods and device and groups of objects recognition methods and device |
| CN109583932A (en) * | 2018-09-29 | 2019-04-05 | 阿里巴巴集团控股有限公司 | A kind of method, apparatus and electronic equipment of generation activity crowd |
| CN110517099A (en) * | 2018-05-22 | 2019-11-29 | 北京京东尚科信息技术有限公司 | Method and apparatus for determining joint supply side |
| WO2020037931A1 (en) * | 2018-08-20 | 2020-02-27 | 平安科技（深圳）有限公司 | Item recommendation method and apparatus, computer device and storage medium |
| CN111695023A (en) * | 2019-03-11 | 2020-09-22 | 北京京东尚科信息技术有限公司 | Information recommendation method and device, storage medium and equipment |
| CN111724183A (en) * | 2019-03-20 | 2020-09-29 | 中国银联股份有限公司 | Merchant recommendation method and merchant recommendation system |
| WO2020252742A1 (en) * | 2019-06-20 | 2020-12-24 | 深圳市云中飞网络科技有限公司 | Resource pushing method and related product |
| CN112288530A (en) * | 2020-10-30 | 2021-01-29 | 广东技术师范大学 | Resource sharing intelligent dining method, device, system, medium and equipment |
| CN112995248A (en) * | 2019-12-12 | 2021-06-18 | 阿里巴巴集团控股有限公司 | Information pushing method, device and equipment |
| CN113051476A (en) * | 2021-03-25 | 2021-06-29 | 北京百度网讯科技有限公司 | Method and apparatus for message transmission |
| CN113254833A (en) * | 2021-06-07 | 2021-08-13 | 深圳市中元产教融合科技有限公司 | Information pushing method and service system based on birth teaching fusion |
| CN114756758A (en) * | 2022-04-29 | 2022-07-15 | 杭州核新软件技术有限公司 | Hybrid recommendation method and system |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060036685A1 (en) * | 2004-07-30 | 2006-02-16 | Microsoft Corporation | Suggesting a discussion group based on indexing of the posts within that discussion group |
| CN104038517A (en) * | 2013-03-05 | 2014-09-10 | 腾讯科技（深圳）有限公司 | Information pushing method based on group relation and server |
| CN104966125A (en) * | 2015-05-06 | 2015-10-07 | 同济大学 | Article scoring and recommending method of social network |
| CN105718579A (en) * | 2016-01-22 | 2016-06-29 | 浙江大学 | Information push method based on internet-surfing log mining and user activity recognition |
| CN106168971A (en) * | 2016-07-08 | 2016-11-30 | 北京麒麟合盛网络技术有限公司 | information subscribing method and device |
-
2017
- 2017-06-29 CN CN201710517520.3A patent/CN107507016A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060036685A1 (en) * | 2004-07-30 | 2006-02-16 | Microsoft Corporation | Suggesting a discussion group based on indexing of the posts within that discussion group |
| CN104038517A (en) * | 2013-03-05 | 2014-09-10 | 腾讯科技（深圳）有限公司 | Information pushing method based on group relation and server |
| CN104966125A (en) * | 2015-05-06 | 2015-10-07 | 同济大学 | Article scoring and recommending method of social network |
| CN105718579A (en) * | 2016-01-22 | 2016-06-29 | 浙江大学 | Information push method based on internet-surfing log mining and user activity recognition |
| CN106168971A (en) * | 2016-07-08 | 2016-11-30 | 北京麒麟合盛网络技术有限公司 | information subscribing method and device |
Cited By (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108537567B (en) * | 2018-03-06 | 2020-08-07 | 阿里巴巴集团控股有限公司 | Method and device for determining target user group |
| CN108537567A (en) * | 2018-03-06 | 2018-09-14 | 阿里巴巴集团控股有限公司 | A kind of determination method and apparatus of targeted user population |
| CN108734348A (en) * | 2018-05-14 | 2018-11-02 | 广东心里程教育集团有限公司 | A kind of method and system of automatic push online course |
| CN110517099A (en) * | 2018-05-22 | 2019-11-29 | 北京京东尚科信息技术有限公司 | Method and apparatus for determining joint supply side |
| CN108898447B (en) * | 2018-07-23 | 2021-09-28 | 北京大前科技有限责任公司 | Method and device for acquiring extended content |
| CN108898447A (en) * | 2018-07-23 | 2018-11-27 | 北京大前科技有限责任公司 | Obtain the method and device of extension content |
| CN109118282A (en) * | 2018-08-08 | 2019-01-01 | 福建百悦信息科技有限公司 | A kind of bimodulus mutual inductance intelligent space user draws a portrait management method and terminal |
| WO2020037931A1 (en) * | 2018-08-20 | 2020-02-27 | 平安科技（深圳）有限公司 | Item recommendation method and apparatus, computer device and storage medium |
| CN109389157A (en) * | 2018-09-14 | 2019-02-26 | 阿里巴巴集团控股有限公司 | A kind of user group recognition methods and device and groups of objects recognition methods and device |
| CN109583932A (en) * | 2018-09-29 | 2019-04-05 | 阿里巴巴集团控股有限公司 | A kind of method, apparatus and electronic equipment of generation activity crowd |
| CN109583932B (en) * | 2018-09-29 | 2023-04-07 | 创新先进技术有限公司 | Method and device for generating active crowd and electronic equipment |
| CN111695023A (en) * | 2019-03-11 | 2020-09-22 | 北京京东尚科信息技术有限公司 | Information recommendation method and device, storage medium and equipment |
| CN111724183A (en) * | 2019-03-20 | 2020-09-29 | 中国银联股份有限公司 | Merchant recommendation method and merchant recommendation system |
| WO2020252742A1 (en) * | 2019-06-20 | 2020-12-24 | 深圳市云中飞网络科技有限公司 | Resource pushing method and related product |
| CN112995248A (en) * | 2019-12-12 | 2021-06-18 | 阿里巴巴集团控股有限公司 | Information pushing method, device and equipment |
| CN112995248B (en) * | 2019-12-12 | 2023-04-07 | 阿里巴巴集团控股有限公司 | Information pushing method, device and equipment |
| CN112288530A (en) * | 2020-10-30 | 2021-01-29 | 广东技术师范大学 | Resource sharing intelligent dining method, device, system, medium and equipment |
| CN112288530B (en) * | 2020-10-30 | 2023-06-02 | 广东技术师范大学 | Resource sharing intelligent dining method, device, system, medium and equipment |
| CN113051476A (en) * | 2021-03-25 | 2021-06-29 | 北京百度网讯科技有限公司 | Method and apparatus for message transmission |
| CN113051476B (en) * | 2021-03-25 | 2023-06-13 | 北京百度网讯科技有限公司 | Method and device for sending message |
| CN113254833A (en) * | 2021-06-07 | 2021-08-13 | 深圳市中元产教融合科技有限公司 | Information pushing method and service system based on birth teaching fusion |
| CN114756758A (en) * | 2022-04-29 | 2022-07-15 | 杭州核新软件技术有限公司 | Hybrid recommendation method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107507016A (en) | A kind of information push method and system | |
| CN108320171B (en) | Hot-sold commodity prediction method, system and device | |
| CN105095219B (en) | Micro-blog recommendation method and terminal | |
| Chen | Multiple criteria ABC inventory classification using two virtual items | |
| CN103295145A (en) | Mobile phone advertising method based on user consumption feature vector | |
| CN102279887A (en) | Method, device and system for classifying documents | |
| CN107274242A (en) | A kind of Method of Commodity Recommendation based on association analysis algorithm | |
| CN109241451B (en) | Content combination recommendation method and device and readable storage medium | |
| CN107423335B (en) | Negative sample selection method for single-class collaborative filtering problem | |
| CN109740072B (en) | Hotel ordering method and system based on POI under OTA platform | |
| CN110990695A (en) | Recommendation system content recall method and device | |
| CN110175895A (en) | A kind of item recommendation method and device | |
| Bahamonde et al. | Power structure in Chilean news media | |
| CN112307762A (en) | Search result sorting method and device, storage medium and electronic device | |
| CN112231593B (en) | Financial information intelligent recommendation system | |
| CN106897419A (en) | The study recommendation method that sorted to level of fusion social information | |
| CN111914159A (en) | Information recommendation method and terminal | |
| CN105574213A (en) | Microblog recommendation method and device based on data mining technology | |
| CN108694211B (en) | Application distribution method and device | |
| CN110825974B (en) | Recommendation system content ordering method and device | |
| CN110852785B (en) | User grading method, device and computer readable storage medium | |
| Gheisari et al. | Data Mining Techniques for Web Mining: A Survey | |
| CN111831898A (en) | Sorting method and device, electronic equipment and readable storage medium | |
| CN108595580B (en) | News recommendation method, device, server and storage medium | |
| US10565603B2 (en) | Segments of contacts |
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 |