CN104298719A - Method and system for conducting user category classification and advertisement putting based on social behavior - Google Patents
Method and system for conducting user category classification and advertisement putting based on social behavior Download PDFInfo
- Publication number
- CN104298719A CN104298719A CN201410492126.5A CN201410492126A CN104298719A CN 104298719 A CN104298719 A CN 104298719A CN 201410492126 A CN201410492126 A CN 201410492126A CN 104298719 A CN104298719 A CN 104298719A
- Authority
- CN
- China
- Prior art keywords
- user
- attributive character
- behavior
- social
- group
- 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
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/955—Retrieval from the web using information identifiers, e.g. uniform resource locators [URL]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
Abstract
The invention discloses a method and system for conducting user category classification and advertisement putting based on social behavior. The method comprises the steps that according to the UID of a user to be classified, social logs corresponding to the record of the UID within a set period are acquired and are used as the user social data of the user; according to the cookie ID corresponding to the UID, access logs corresponding to the record of the cookie ID within the set period are acquired and are used as the user behavior data of the user; user behavior property characters are extracted from the user behavior data, and user social property characters are extracted from the user social data; the user is classified to the corresponding category according to the extracted user behavior property characters and the extracted user social property characters. According to the method and system, the accuracy of category classification of the user and the effectiveness of advertisement putting can be improved.
Description
Technical field
The present invention relates to internet arena, particularly relate to and a kind ofly carry out the category division of user, advertisement placement method and system based on Social behaviors.
Background technology
Along with the development of Internet technology, the advantage that Internet advertising is quick with it, convenient, dirigibility is strong instead of rapidly traditional media advertisement.Internet advertising throws in advertisement by web advertisement platform exactly on network, utilize the ad banner on website, text link, multimedia method, in internet publication or releasing advertisements, by a kind of high-tech advertising campaign mode of network delivery to Internet user.Compared with the outdoor advertising of showing appreciation for somebody with traditional four large propagation media (newspaper, magazine, TV, broadcast) advertisement and enjoying, Internet advertising take internet as advertising media, having advantageous advantage, is the important part implementing Modern Marketing media strategy.
And along with the high speed development of internet advertisement technology, the needs of advertiser to target audience are also more and more higher.And traditional media advertisement not to be life cycle shorter, it is complicated to make, be exactly that dirigibility and promptness are poor, be also very easily subject to the interference of extraneous factor simultaneously; More mainly they cannot audient crowd required for accurate lock oneself, therefore need badly provide a kind of can towards crowd, in order to promote the advertisement mode that promotion effect promotes greatly.
In practical application, IGRP (Internet Gross Rating Points, total internet rating point) be exactly scheme is sold in the advertisement towards crowd that a kind of GRP from traditional tv advertisement of advertisement (Gross Rating Point, total rating point) develops and comes.Deeply can be excavated the feature such as age, sex, interest of audient user by the large data edge in internet, then according to the feature excavated, category division be carried out to user, so that advertiser carries out advertisement putting targetedly, promote promotion effect.
Existingly provide a kind of class of subscriber division methods, user mainly according to the access log of user in advertisement putting website, is divided into corresponding classification by it.Particularly, acquisition user after the access log of advertisement putting website, the page of user's access can be analyzed, obtain the feature of the page of user's access; Then, according to the feature of the page of user's access, the classification of user is inferred.But, in fact, according to the anti-classification pushing away user of feature of the page that one-side user accesses, its confidence level is not high, cause the accuracy of the category division result of user not high, then cause audient's less pertinence of the advertisement of throwing according to category division result, promotion effect poor, reduce the validity of advertisement putting.
Therefore, be necessary to provide a kind of class of subscriber division methods improving category division accuracy.
Summary of the invention
For the defect that above-mentioned prior art exists, the invention provides and a kind ofly carry out the category division of user, advertisement placement method and system, in order to improve accuracy and the advertisement putting validity of the category division of user based on Social behaviors.
According to an aspect of the present invention, provide a kind of category classification method carrying out user based on Social behaviors, comprising:
Obtain the user behavior data of user to be sorted in setting cycle, and user social contact data;
From described user behavior data, extract user behavior attributive character, from described user social contact data, extract user social contact attributive character;
According to the user behavior attributive character extracted and user social contact attributive character, described user is divided into respective classes.
Preferably, the described user behavior attributive character according to extraction and user social contact attributive character, be divided into respective classes by described user, specifically comprise:
By the user behavior attributive character of described user, mate with the user behavior attributive character recorded in the behavioural characteristic rule base built in advance, judge the behavior group belonging to described user according to matching result; Wherein, for Mei Zhong behavior group in described behavioural characteristic rule base, have recorded the behavior user behavior attributive character that has of group;
By the user social contact attributive character of described user, mate with the user social contact attributive character recorded in the social characteristics rule base built in advance, judge the social group belonging to described user according to matching result; Wherein, for often kind of social group in described social characteristics rule base, have recorded the user social contact attributive character that this social group has;
Find out group common between behavior group belonging to described user and social group, and described user is divided in the common group found out.
Preferably, the described user behavior attributive character according to extraction and user social contact attributive character, be divided into respective classes by described user, specifically comprise:
The user behavior attributive character of the described user extracted and user social contact attributive character are merged and gathers the user characteristics obtaining described user;
The user characteristics of the described user obtained is input in the category division model for category division; By described category division model, described user is divided into respective classes.
Preferably, described category division model is training in advance:
For each training user selected from the user of website, after the user behavior attributive character of adding up this training user and user social contact attributive character, merge and gather the user characteristics obtaining this training user; And determine the behavior group belonging to this training user according to the user behavior attributive character of this training user, the social group belonging to this training user is determined according to the user social contact attributive character of this training user, and using group common between the behavior group belonging to this training user and social group as group belonging to this training user;
Group, user characteristics belonging to each training user, as training data, are used described training data to use preset many labels multi-classification algorithm to carry out model training, obtain category division model.
Preferably, described obtain category division model before, also comprise:
For each test subscriber selected from the user of website, after the user behavior attributive character of adding up this test subscriber and user social contact attributive character, merge and gather the user characteristics obtaining this test subscriber; And determine the behavior group belonging to this test subscriber according to the user behavior attributive character of this test subscriber, the social group belonging to this test subscriber is determined according to the user social contact attributive character of this test subscriber, and using group common between the behavior group belonging to this test subscriber and social group as group belonging to this test subscriber;
Using group, user characteristics belonging to each test subscriber as test data; And
When using described training data to use preset many labels multi-classification algorithm to carry out model training, also using described test data to use preset many labels multi-classification algorithm to carry out model training, obtaining described category division model.
Present invention also offers a kind of advertisement placement method, comprising:
After website receives the request of access of user, the user identity of the identification number cookieID and correspondence that obtain described user proves UID;
According to cookieID and UID of described user, adopt as arbitrary in claim 1-5 as described in method, obtain the category division result of described user;
Using the category division result that obtains as search keyword, find out the ad content matched with described search keyword;
The ad content found out is play in the advertisement played column of the webpage returned to described user for described request of access.
According to another aspect of the present invention, additionally provide a kind of category division system of carrying out user based on Social behaviors, comprising:
User data acquisition module, for obtaining the user behavior data of user to be sorted in setting cycle and user social contact data and exporting;
Attributive character abstraction module, for extracting user behavior attributive character in the user behavior data from described user data acquisition module output, extracts user social contact attributive character from the user social contact data that described user data acquisition module exports;
Class of subscriber divides module, for the user behavior attributive character that extracts according to described attributive character abstraction module and user social contact attributive character, described user is divided into respective classes.
Preferably, described class of subscriber division module specifically comprises: first user category division unit or the second class of subscriber division unit; Wherein,
Described first user category division unit specifically comprises:
Behavioural characteristic coupling subelement, for the user behavior attributive character that described attributive character abstraction module is extracted, mate with the user behavior attributive character recorded in the behavioural characteristic rule base built in advance, judge the behavior group belonging to described user according to matching result; Wherein, for Mei Zhong behavior group in described behavioural characteristic rule base, have recorded the behavior user behavior attributive character that has of group;
Social characteristics coupling subelement, for the user social contact attributive character that described attributive character abstraction module is extracted, mate with the user social contact attributive character recorded in the social characteristics rule base built in advance, judge the social group belonging to described user according to matching result; Wherein, for often kind of social group in described social characteristics rule base, have recorded the user social contact attributive character that this social group has;
Category division subelement, for receive described behavioural characteristic matching unit export user belonging to behavior group and described social characteristics matching unit export user belonging to social group; Find out group common between behavior group belonging to described user and social group, and described user is divided in the common group found out; And
Described second class of subscriber division unit is used for user behavior attributive character and the user social contact attributive character of the described user extracted by described attributive character abstraction module, merges and gathers the user characteristics obtaining described user; And the user characteristics of the described user obtained is input in the category division model for category division; By described category division model, described user is divided into respective classes.
Preferably, the category division system of carrying out user based on Social behaviors also comprises:
Partitioning model training module, for for each training user selected from the user of website, after the user behavior attributive character of adding up this training user and user social contact attributive character, merges and gathers the user characteristics obtaining this training user; And determine the behavior group belonging to this training user according to the user behavior attributive character of this training user, the social group belonging to this training user is determined according to the user social contact attributive character of this training user, and using group common between the behavior group belonging to this training user and social group as group belonging to this training user; Group, user characteristics belonging to each training user, as training data, are used described training data to use preset many labels multi-classification algorithm to carry out model training, obtain category division model.
Present invention also offers a kind of advertisement delivery system, comprising:
The above-mentioned user data acquisition module carried out based on Social behaviors in the category division system of user, attributive character abstraction module, class of subscriber divide module;
Web-page requests receiver module, for receive user request of access after, obtain the cookieID of user and the UID of correspondence, and be sent to described user data acquisition module;
Advertising inquiry module, for described class of subscriber being divided the category division result of the user that module exports as search keyword, finds out the ad content matched with described search keyword;
Request processing module, after inserting ad content that described advertising inquiry module searches goes out in the advertisement played column of webpage of asking in described request of access, returns described webpage to described user;
Described user data acquisition module specifically for according to the cookieID of user and the UID of correspondence, obtains the user behavior data of user in setting cycle and user social contact data and exports.
In technical scheme of the present invention, according to the user behavior data of user on website, user social contact data under social platform, extract user behavior attributive character and user social contact attributive character respectively; Then, the category division of user is carried out according to the user behavior attributive character extracted and user social contact attributive character.Compare existing only according to the anti-interest pushing away user of the information of user to access pages, technical scheme provided by the invention, consider user behavior attributive character and user social contact attributive character to carry out class of subscriber division, add the feature rich degree of category division, and improve class of subscriber division accuracy, also just can improve crowd's positional accuracy; Like this, based on higher crowd's positional accuracy, the specific aim of the audient of advertisement, promotional videos etc. can be improved, throw in validity.
Accompanying drawing explanation
Fig. 1 is the schematic flow sheet carrying out the category classification method of user based on Social behaviors of the embodiment of the present invention;
Fig. 2 is the schematic flow sheet of the advertisement placement method of the embodiment of the present invention;
Fig. 3 is the schematic flow sheet extracting user behavior attributive character of the embodiment of the present invention;
Fig. 4 is the schematic flow sheet extracting user social contact attributive character of the embodiment of the present invention;
Fig. 5 is the training method schematic flow sheet of the category division model of the embodiment of the present invention;
Fig. 6 is the structural representation carrying out the category division system of user based on Social behaviors of the embodiment of the present invention;
Fig. 7 is the structural representation of the first user category division unit of the embodiment of the present invention;
Fig. 8 is the structural representation of the advertisement delivery system of the embodiment of the present invention.
Embodiment
Carry out clear, complete description below with reference to accompanying drawing to technical scheme of the present invention, obviously, described embodiment is only a part of embodiment of the present invention, instead of whole embodiments.Based on the embodiment in the present invention, other embodiments all that those of ordinary skill in the art obtain under the prerequisite not making creative work, all belong to the scope that the present invention protects.
The term such as " module " used in this application, " system " is intended to comprise the entity relevant to computing machine, such as but not limited to hardware, firmware, combination thereof, software or executory software.Such as, module can be, but be not limited in: the thread of the process that processor runs, processor, object, executable program, execution, program and/or computing machine.For example, application program computing equipment run and this computing equipment can be modules.One or more module can be positioned at an executory process and/or thread, and module also and/or can be distributed on a computing machine between two or more platform computing machines.
The present inventor finds, for in website, situation except user's accession page on the web site can reflect the interest of user to a certain extent, and the blog article that user issues in the social platform of this website log, that forward, point is praised also can react the interest of user to a certain extent.
Therefore, the present inventor considers, can using the access log of user on website as user behavior data, using the social daily record of user under social platform as user social contact data, and extract user behavior attributive character and user social contact attributive character respectively from user behavior data, user social contact data; Then, the category division of user can be carried out according to the user behavior attributive character extracted and user social contact attributive character.Compare existing only according to the anti-interest pushing away user of the information of user to access pages, technical scheme provided by the invention, consider user behavior attributive character and user social contact attributive character to carry out class of subscriber division, the feature rich degree of category division can be increased, and improve class of subscriber division accuracy, also just can improve crowd's positional accuracy; Based on higher crowd's positional accuracy, the specific aim of the audient of advertisement, promotional videos etc. can be improved, throw in validity.
Technical scheme of the present invention is described in detail below in conjunction with accompanying drawing.
Embodiments provide a kind of category classification method carrying out user based on Social behaviors, its flow process as shown in Figure 1, specifically can comprise the steps:
S101: the user behavior data obtaining user to be sorted in setting cycle, and user social contact data.
In practical application, user is when sending request of access to website, website usually can be passed through cookie ID (Identity, identification number) and mark user, and records all access logs of this user.Correspondingly, if user registers in the social platform of this website, unique UID (User Identification, user identity proves) can be obtained; Like this, when this user logs in social platform by UID at every turn, the cookie ID of this user and all social daily records to should UID, can be recorded in website.Social platform in the embodiment of the present invention can be specifically microblogging, push away top grade.
Therefore, particularly, then after the UID determining user to be sorted, can according to the UID of this user, can obtain should social daily record in the setting cycle of UID record, and as the user social contact data of this user.Further, can cookie ID corresponding to this UID, obtain should access log in the setting cycle that records of cookie ID, and as the user behavior data of this user.
Wherein, the access log of user in setting cycle specifically can comprise: each advertisement and number of clicks etc. thereof of each page that user accesses in setting cycle on website and access times thereof, click.The social daily record of user in setting cycle specifically can comprise: the blog article of user, the log-on message of user, the concern information of user, LBS (Location Based Service, the position-based service) positional information etc. of user.
S102: extract user behavior attributive character from user behavior data, extracts user social contact attributive character from user social contact data.
Particularly, according to the user behavior data obtained, can therefrom extract the user behavior attributive character of user, such as, the sex, age, interest etc. of user.Further, according to the user social contact data obtained, can therefrom extract the user social contact attributive character of user, such as, sex, age, blog article feature, concern user characteristics, LBS feature etc.In practical application, the user behavior attributive character extracted from user behavior data also can be called the feature based on cookie ID; The user social contact attributive character extracted from user social contact data also can be called the feature based on UID.
About how extracting above-mentioned user behavior attributive character from user behavior data, and how from user social contact data, to extract above-mentioned user social contact attributive character, will in follow-up detailed introduction.
S103: according to the user behavior attributive character extracted and user social contact attributive character, user is divided into respective classes.
Particularly, by the user behavior attributive character of user, can mate with the user behavior attributive character recorded in the behavioural characteristic rule base built in advance, judge the behavior group belonging to user according to matching result.Wherein, for Mei Zhong behavior group in behavioural characteristic rule base, have recorded the behavior user behavior attributive character that has of group.Correspondingly, by the user social contact attributive character of the user of acquisition, can mate with the user social contact attributive character recorded in the social characteristics rule base built in advance, judge the social group belonging to user according to matching result.Wherein, for often kind of social group in social characteristics rule base, have recorded the user social contact attributive character that this social group has.
In the embodiment of the present invention, behavior group in behavioural characteristic rule base and the social group in social characteristics rule base all pre-set, such as, can comprise: Investment & Financing group, sport and body-building group, city white collar group, group of parent-offspring family, the youth campus group, video display group, photography group etc.
If consider that user is interested in certain class things, it is on access navigation patterns and Social behaviors can show similarity usually.Therefore, group common between the behavior group belonging to user and social group can be searched, if group common between the behavior group found out belonging to user and social group, then user can be divided in the common group found out.Compare the existing classification divided based on one-side behavior, in the solution of the present invention, group common between the behavior group judged respectively by the feature based on cookie ID and the feature based on UID and social group is as the category division result of user, and accuracy is higher.
In practical application, can for the Mei Zhong behavior group in behavioural characteristic rule base, the user behavior attributive character extracted is mated with the user behavior attributive character that behavior group has, if the match is successful, then can judge the behavior group of behavior group belonging to this user; If do not mate, then can judge that this user does not belong to behavior group.Correspondingly, can for often kind in social characteristics rule base social group, the user social contact attributive character extracted is mated with the user social contact attributive character that this social group has, if the match is successful, then can judge the social group of this social group belonging to this user; If do not mate, then can judge that this user does not belong to this social group.
Further, for Mei Zhong behavior group in behavior characterization rules storehouse, the behavior, the user behavior attributive character that has of group specifically can comprise several prerequisite user behavior attributive character of user belonging to behavior group, also can comprise the user behavior attributive character that some users belonging to behavior group generally possess.
Therefore, when the user behavior attributive character extracted is mated with the user behavior attributive character that behavior group has, if the user behavior attributive character extracted includes the prerequisite user behavior attributive character of user belonging to behavior group, then can judge that the match is successful; Otherwise, judge that coupling is unsuccessful.Correspondingly, for often kind of social group in social characteristics rule base, the user social contact attributive character that this social group has specifically can comprise several prerequisite user social contact attributive character of user belonging to this social group, also can comprise the user social contact attributive character that some users belonging to this social group generally possess.Therefore, when the user social contact attributive character extracted is mated with the user social contact attributive character that this social group has, if the user social contact attributive character extracted includes the prerequisite user social contact attributive character of user belonging to this social group, then can judge that the match is successful; Otherwise, judge that coupling is unsuccessful.
In addition, if there is not common group between the behavior group belonging to user and social group, then the user behavior attributive character of the user extracted and user social contact attributive character can be merged and gather, obtain the user characteristics of user.Then, according to the user characteristics of the user obtained and the category division model of training in advance, user is divided into respective classes.Particularly, multiple Fusion Features that can be adopted are nearly a user characteristics; To the feature of nearly justice be there is no separately as a user characteristics; According to the fusion rule preset, be a user characteristics by multiple Fusion Features of contradiction.Such as, if the age characteristics in the user behavior attributive character extracted contradicts with the age characteristics in the user social contact attributive character extracted, then can according to preset the fusion rule about the age, using the age characteristics in user social contact attributive character directly as the age characteristics in user characteristics.
Present invention also offers a kind of more excellent embodiment, carrying out step S102: from user behavior data, extract user behavior attributive character, extract user social contact attributive character from user social contact data after, directly the user behavior attributive character of the user of extraction and user social contact attributive character can be carried out fusion to gather, obtain the user characteristics of user; Then, what the user characteristics of the user obtained can be input to training in advance is used in the category division model of category division; By category division model, user is divided into respective classes.
In the embodiment of the present invention, about the training of category division model, will in follow-up detailed introduction.
In practical application, according to extract user behavior attributive character and user social contact attributive character, after user is divided into respective classes, the classification that user can be divided into stores as category division result.Consider that in practical application, this user is in different computer log websites, website to should the cookie ID of user record different, that is, may there is multiple cookie ID in corresponding same user; But the UID of same user's existence anduniquess; Therefore, further, by the UID corresponding stored of the category division result of user and user, follow-up use can be convenient to.Certainly, in practical application, also can by cookie ID and the UID of user and the common corresponding stored of category division result.
Based on above-mentioned category classification method, present invention also offers a kind of advertisement placement method, its flow process as shown in Figure 2, specifically can comprise the steps:
S201: after website receives the request of access of user, obtains the cookie ID of user and the UID of correspondence.
Particularly, after website receives the request of access of user, except can normally obtain the webpage corresponding with request of access; Can also can obtain the cookie ID of this user and the UID of correspondence thereof.
S202: according to cookie ID and UID obtained, obtains the user behavior data of user in setting cycle, and user social contact data.
S203: extract user behavior attributive character from the user behavior data obtained, extracts user social contact attributive character from the user social contact data obtained; And according to the user behavior attributive character extracted and user social contact attributive character, user is divided into respective classes, obtains the category division result of user.
More preferably, website receives the request of access of user, after obtaining cookie ID and UID of user, directly can find out the category division result corresponding with cookie ID and UID obtained from class of subscriber division result storehouse, as the category division result of this user.Wherein, corresponding each user in class of subscriber division result storehouse, stores the category division result and cookie ID and UID corresponding with it that mark off according to the method as step S101-S103.
S204: using the category division result that obtains as search keyword, finds out and the ad content searched for keyword and match; The ad content found out is play in the advertisement played column of the webpage returned to user for request of access.
Particularly, method well-known to those skilled in the art can be adopted, after acquisition search keyword, can find out from backstage, website and the ad content searched for keyword and match; And the ad content of all acquisitions is joined in the advertisement played column of the webpage returned to user for request of access.Like this, when the webpage of user in access websites, website can push user according to the category division result of user to user may interested advertisement, improves the precision of advertisement putting.
In the embodiment of the present invention, before category division is carried out to user, can in advance for each page of website, according to the content of this page and the characteristic of audient user, corresponding stored has in each page properties feature of each page properties and score thereof.
Wherein, page properties specifically can comprise: age, sex and hobby; There is under each page properties different page properties features.Like this, for each page, can according to the characteristic of the audient user of this page, this page has corresponding score in different page properties features.For ease of describing, in the embodiment of the present invention, the score of page properties feature can also be belonged to feature referred to as page and dividing.
In practical application, the page properties feature at page properties-age (also can referred to as age characteristics) specifically can comprise: 10 ~ 20 years old or 21 ~ 30 years old or 31 ~ 40 years old or be more than or equal to 41 years old.Such as, for a certain page of website, page properties feature corresponding to this page-10 ~ 20 years old must be divided into 0.1; Page properties feature corresponding to this page-21 ~ 30 years old must be divided into 0.7; Page properties feature corresponding to this page-31 ~ 40 years old must be divided into 0.1; The page properties feature that this page is corresponding-be more than or equal to 41 years old must be divided into 0.1.
The page properties feature of page properties-sex (also can referred to as sex character) specifically can comprise: the male sex, women.Such as, for a certain page of website, website, page properties feature-male sex that this page is corresponding must be divided into 0.3; Page properties feature-women that this page is corresponding must be divided into 0.7.
The page properties feature of page properties-interest (also can referred to as interest characteristics) specifically can comprise; Reading, tourism, house property, second-hand house, basketball, football, financing etc.Such as, for a certain page of website, website, page properties feature-financing that this page is corresponding must be divided into 0.5; Page properties feature-reading that this page is corresponding must be divided into 0.3; Page properties feature-house property that this page is corresponding must be divided into 0.2; And the score of other page properties features under page properties-interest corresponding to this page is 0.
The each page properties feature corresponding based on the above-mentioned each page prestored in advance and score thereof, how from user behavior data, user behavior attributive character is extracted about what mention in above-mentioned steps S102, as shown in Figure 3, specifically can extract as follows:
S301: from the user behavior data obtained, determine each page that user accesses in setting cycle.
Particularly, except each page that user accesses in setting cycle can be determined from user behavior data, the access times of each page can also be determined.
S302: each page of accessing in setting cycle for user, obtains each page feature of this page under each page properties and score thereof.
S303: according to the score of all page properties features of each page under each page properties, determine the page feature of user under each page properties, and using the page feature determined as the user behavior attributive character extracted.
Particularly, for page properties-age, for each page properties feature (age characteristics) under this page properties, the score of this page properties feature that can store corresponding to each page obtained, calculates the mean value of the score of this page properties feature; Choose the highest page properties feature of mean value as the page feature of user under this page properties, and as the user behavior attributive character extracted.Correspondingly, for page properties-sex, for each page properties feature (sex character) under this page properties, the score of this page properties feature that can store corresponding to each page obtained, calculates the mean value of the score of this page properties feature; Choose the highest page properties feature of mean value as the page feature of user under this page properties, and as the user behavior attributive character extracted.
Further, for page properties-interest, after obtaining each page properties feature under page properties-interest and score thereof stored corresponding to each page of user's access, for each page properties feature under page properties-interest, the score of this page properties feature of each page corresponding stored is added up, if the cumulative score of this page properties feature is greater than the page properties score threshold of setting, then using this page properties feature as the page feature of this user under page properties-interest, and as the user behavior attributive character extracted.Wherein, page properties score threshold specifically can be set to 0.
In practical application, relative to the interest of user, the sex and age of user belongs to the base attribute of user, therefore, in the embodiment of the present invention, the user behavior attributive character of user can be divided into: the base attribute feature of Behavior-based control, the interest characteristics of Behavior-based control.Like this, the page feature of user under page properties-age and the page feature under page properties-sex specifically belong to the base attribute feature of the Behavior-based control in the user behavior attributive character of user; The page feature of user under page properties-interest specifically belongs to the interest characteristics of Behavior-based control in the user behavior attributive character of user.
In practical application, if a certain channel that user classifies in advance to website is interested, the access times of its page under this channel will be many; Therefore, more preferably, in the embodiment of the present invention, while carrying out step S303, user behavior attributive character can also be extracted as follows:
S304: the channel belonging to each page that user accesses in setting cycle, counts each access channel of user on website, and using each access channel of counting as the user behavior attributive character extracted.
Particularly, for each page that user accesses in setting cycle, the channel belonging to this page is determined; According to the channel belonging to each page determined, count all access channels on the inherent website of user's setting cycle, and using the access channel that counts as the one in the interest characteristics of Behavior-based control in the user behavior attributive character of user.Wherein, on website, the subordinate channel of each page divides in advance.
Further, for each access channel counted, the feature score of this access channel can be determined according to the inherent access times belonging to each page of this access channel of user's setting cycle.Like this, more preferably, the access channel that feature score does not exceed the access channels feature score threshold of setting can be rejected from the behavior property feature of user, ensure the accuracy of the category division result divided based on user behavior attributive character.
Such as, for each access channel, user's setting cycle inherence can be belonged to the feature score of summation as this channels feature of the access times of each page of this access channel.Or, also for each access channel, user's setting cycle inherence can be belonged to the visitation frequency of access times summation as this access channel of each page of this access channel; And using the feature score of the ratio of the visitation frequency summation of the visitation frequency of this channels feature and all access channels counted as this channels feature.
In practical application, except each page that user accesses in setting cycle and access times thereof can be determined from user behavior data, each advertisement that user clicks in setting cycle and number of clicks thereof can also be determined.
Therefore, more preferably, in the embodiment of the present invention, while carrying out step S303, user behavior attributive character can also be extracted as follows:
S305: the industry belonging to each advertisement that user clicks in setting cycle on website, counts each click industry of user on website, and using the click industry counted as the user behavior attributive character extracted.
Particularly, for each advertisement that user clicks in setting cycle on website, the industry belonging to this advertisement is determined; According to the industry belonging to each advertisement determined, count all click industries on the inherent website of user's setting cycle, and using the click industry that counts as the one in the interest characteristics of Behavior-based control in the user behavior attributive character of user.Wherein, the industry on website belonging to each advertisement divides in advance.
Further, for each click industry counted, the number of clicks of each advertisement of this click industry can be belonged to according to user's setting cycle inherence, determine the feature score of this click industry.Like this, more preferably, the click industry that feature score does not exceed the click industrial characteristic score threshold of setting can be rejected from the user behavior attributive character of user, ensure the accuracy of the category division result that user behavior attributive character divides.
Such as, for each click industry, user's setting cycle inherence can be belonged to the feature score of number of clicks summation as this click industry of each advertisement of this click industry; Or, also can for each click industry, user's setting cycle inherence is belonged to the click frequency of number of clicks summation as this click industry of each advertisement of this click industry, and using the feature score of the ratio of the click frequency summation of the click frequency of this click industry and all click industries as this click industry.
How from user social contact data, to extract user social contact attributive character about what mention in above-mentioned steps S102, as shown in Figure 4, specifically can extract as follows:
S401: from the user social contact data obtained, determine user's every section of blog article in setting cycle.
Wherein, the blog article of user specifically can comprise: user initiates, forward, comment on, put the blog article of praising.
S402: for the every section of blog article determined, extracts the blog article descriptor of this blog article; And using the blog article descriptor of all blog articles that the extracts user social contact attributive character as user.
In the embodiment of the present invention, the user social contact attributive character of user can be divided into: based on the base attribute feature of social activity, the interest characteristics based on social activity.Wherein, the blog article descriptor extracted specifically belongs in the user social contact attributive character of user based on the one in the interest characteristics of social activity.About the blog article descriptor how extracted in blog article, the common technology means of those skilled in the art can be adopted, be not described in detail in this.
Further, for each blog article descriptor extracted, according to the occurrence number of this blog article descriptor in all blog articles, the feature score of this blog article descriptor can also be determined.In practical application, for each blog article descriptor, can using the occurrence number of this blog article descriptor in all blog articles that count directly as the feature score of this blog article descriptor.Or, also can using the occurrence number of this blog article descriptor in all blog articles and each blog article descriptor the ratio of the occurrence number summation in all blog articles as the feature score of this blog article descriptor.Like this, more preferably, the blog article descriptor that feature score does not exceed the blog article descriptor feature score threshold value of setting can be rejected from the user social contact attributive character of user, ensure the accuracy of the category division result divided based on user social contact attributive character.
In practical application, need registration before the social platform of user in website logs in, and the log-on message of user can comprise usually: the age, sex, educational background, occupational information etc. of user.
Therefore, more preferably, in the embodiment of the present invention, when carrying out step S401, S402, user social contact attributive character can also be extracted as follows:
S403: the log-on message determining user from the user social contact data obtained; And from the log-on message of user, extract the user social contact attributive character of user.
Particularly, can using the information such as age, sex, educational background of user that extracts from the log-on message of user directly as the base attribute feature based on social activity in the user social contact attributive character of user.
Further, in social platform, user according to the situation of oneself, can select customized label.Consider that customized label can reflect the hobby of user to a certain extent.Therefore, further, in the embodiment of the present invention, while carrying out step S401, S402, user social contact attributive character can also be extracted as follows:
S404: the customized label determining user from the user social contact data obtained, and using the customized label determined as the user social contact attributive character extracted.
Particularly, can using the customized label of user determined from the user social contact data obtained in the user social contact attributive character of user based on the interest characteristics of social activity.
In practical application, user usually can pay close attention to other users in social platform, and the object that user pays close attention to can reflect the interest of user usually to a certain extent.Therefore, more preferably, further, in the embodiment of the present invention, while carrying out step S401, S402, the user social contact attributive character of user can also be extracted according to the concern information of user, specifically can extract user social contact attributive character as follows:
The social daily record of user in setting cycle specifically can comprise: the blog article of user, the log-on message of user, the concern information of user, LBS (Location Based Service, the position-based service) positional information etc. of user.
S405: determine the concern information of user from the user social contact data obtained, and the authenticated belonging to specific industry extracting that from the concern information determined user pays close attention to; For each authenticated extracted, using the industry belonging to this authenticated as the user social contact attributive character extracted.
Particularly, can using user pay close attention to each authenticated belonging to industry in the user social contact attributive character of this user based on the interest characteristics of social activity.
Such as, if user has paid close attention to@appoint the authenticated that will is strong ,@Sina house property ,@Beijing self-housing etc. belong to real estate industry, then can using real estate industry as the user social contact attributive character extracted.
Further, for each authenticated extracted, after determining the industry belonging to this authenticated, according to industry belonging to each authenticated, the industry that user pays close attention to can be counted; For the industry that each user counted pays close attention to, according to the quantity belonging to the authenticated of the sector in the concern information of user, the feature score of the sector can also be calculated.Like this, more preferably, the industry that feature score does not exceed the industrial characteristic score threshold of setting can be rejected from the user social contact attributive character of user, ensure the accuracy of the category division result divided based on user social contact attributive character.
In practical application, the feature of some geographic area is obvious (characteristic area), such as, and National Stadium, region, tourist attractions, commerce and trade district etc.; The place if user often comes in and goes out, can reflect the interest of user to a certain extent.Therefore, Region dividing can be carried out in advance, these be had the region of obvious characteristic as characteristic area, and to should the interest that is associated of characteristic area storage.
The present inventor considers, the user of website often logs in social platform based on certain factor (such as: work, interest etc.) through a certain position of being everlasting; Therefore, can consider, according to the LBS position of user and the characteristic area that divides in advance, to extract user social contact attributive character.
Particularly, because the LBS service of social platform can the login position of recording user, therefore, from social daily record, the LBS position in user's setting cycle to be positioned can directly be obtained.And then, for each LBS position obtained, according to the characteristic area belonging to this LBS position, using the interest characteristics based on social activity in the user social contact attributive character of user of the interest associated by this characteristic area.
In the embodiment of the present invention, be training in advance about the category division model mentioned in step S103, as shown in Figure 5, specifically can train as follows:
S501: for each training user selected from the user of website, after the user behavior attributive character of adding up this training user and user social contact attributive character, merges and gathers the user characteristics obtaining this training user.
S502: for each training user, determine the behavior group belonging to this training user according to the user behavior attributive character of this training user.
S503: for each training user, determine the social group belonging to this training user according to the user social contact attributive character of this training user.
S504: for each training user, using group common between the behavior group belonging to this training user and social group as group belonging to this training user.
S505: using group, user characteristics belonging to each training user as training data, uses training data to use preset many labels multi-classification algorithm to carry out model training, obtains category division model.
Wherein, preset many labels multi-classification algorithm is specifically as follows M3l algorithm or Multiboost algorithm.
In the embodiment of the present invention, step S501-S504 specifically with reference to the step S101-S103 in the above-mentioned category classification method carrying out user based on Social behaviors, can be not described in detail in this.
More preferably, in order to ensure to obtain the higher category division model of degree of accuracy, improve the accuracy of the category division result of user, when using training data to use preset many labels multi-classification algorithm to carry out model training, can also test data can be used further to use preset many labels multi-classification algorithm to carry out model training and obtain category division model.In practical application, the division effect (such as degree of accuracy and recall rate) of test data to category division model can be used to assess, obtain the category division model of effect optimum with this.Wherein, test data specifically can obtain in the following way:
For each test subscriber selected from the user of website, after the user behavior attributive character of adding up this test subscriber and user social contact attributive character, merge and gather the user characteristics obtaining this test subscriber; And determine the behavior group belonging to this test subscriber according to the user behavior attributive character of this test subscriber, the social group belonging to this test subscriber is determined according to the user social contact attributive character of this test subscriber, and using group common between the behavior group belonging to this test subscriber and social group as group belonging to this test subscriber; Using group, user characteristics belonging to each test subscriber as test data.
In practical application, in order to improve model training efficiency, pre-service can be carried out for the training data of all training users, specifically can comprise: the operations such as coding, normalization, dimensionality reduction.
According to the above-mentioned category classification method carrying out user based on Social behaviors, the embodiment of the present invention additionally provides a kind of category division system of carrying out user based on Social behaviors, as shown in Figure 6, specifically can comprise: user data acquisition module 601, attributive character abstraction module 602 and class of subscriber divide module 603.
Wherein, user data acquisition module 601 is for obtaining the user behavior data of user to be sorted in setting cycle and user social contact data and exporting.
Particularly, user data acquisition module 601 can according to the UID of user to be sorted, obtains should social daily record in the setting cycle of UID record, and as the user social contact data of this user.Further, can cookie ID corresponding to this UID, obtain should access log in the setting cycle that records of cookie ID, and as the user behavior data of this user.
Attributive character abstraction module 602 extracts user behavior attributive character in the user behavior data that exports from user data acquisition module 601, from the user social contact data that user data acquisition module 601 exports, extract user social contact attributive character.
Class of subscriber divides module 603 for the user behavior attributive character that extracts according to attributive character abstraction module 602 and user social contact attributive character, and user is divided into respective classes.
More preferably, in the embodiment of the present invention, the category division system of carrying out user based on Social behaviors also comprises further: partitioning model training module 604.
Partitioning model training module 604 for for each training user selected from the user of website, after the user behavior attributive character of adding up this training user and user social contact attributive character, merges and gathers the user characteristics obtaining this training user; And determine the behavior group belonging to this training user according to the user behavior attributive character of this training user, the social group belonging to this training user is determined according to the user social contact attributive character of this training user, and using group common between the behavior group belonging to this training user and social group as group belonging to this training user; Using group, user characteristics belonging to each training user as training data, use training data to use preset many labels multi-classification algorithm to carry out model training, obtain category division model.
In the embodiment of the present invention, class of subscriber divides module 603 and specifically can comprise: first user category division unit or the second class of subscriber division unit.
Wherein, first user category division unit is used for the user behavior attributive character, user social contact attributive character, the behavioural characteristic rule base built in advance and the social characteristics rule base that extract according to attributive character abstraction module 602, and user is divided into respective classes.
Second class of subscriber division unit is used for user behavior attributive character and the user social contact attributive character of the user extracted by attributive character abstraction module 602, merges and gathers the user characteristics obtaining user; And the user characteristics of the user obtained is input in the category division model for category division; By category division model, user is divided into respective classes.
Further, as shown in Figure 7, first user category division unit can specifically comprise: behavioural characteristic coupling subelement 701, social characteristics coupling subelement 702 and category division subelement 703.
The user behavior attributive character of behavioural characteristic coupling subelement 701 for attributive character abstraction module 602 is extracted, mate with the user behavior attributive character recorded in the behavioural characteristic rule base built in advance, judge the behavior group belonging to described user according to matching result.Wherein, for Mei Zhong behavior group in behavioural characteristic rule base, have recorded the behavior user behavior attributive character that has of group.
The user social contact attributive character of social characteristics coupling subelement 702 for attributive character abstraction module 602 is extracted, mate with the user social contact attributive character recorded in the social characteristics rule base built in advance, judge the social group belonging to described user according to matching result.Wherein, for often kind of social group in social characteristics rule base, have recorded the user social contact attributive character that this social group has.
The social group belonging to user that behavior group belonging to the user that category division subelement 703 exports for reception behavior features characteristic matching subelement 701 and social characteristics coupling subelement 702 export; Find out group common between behavior group belonging to user and social group, and user is divided in the common group found out.
In the embodiment of the present invention, based on above-mentioned advertisement placement method and class of subscriber dividing system, the embodiment of the present invention additionally provides a kind of advertisement delivery system, as shown in Figure 8, specifically can comprise: carry out the user data acquisition module 601 in the category division system of user, attributive character abstraction module 602, class of subscriber division module 603, web-page requests receiver module 801, advertising inquiry module 802 and request processing module 803 based on Social behaviors.
Wherein, web-page requests receiver module 801 for receive user request of access after, obtain the cookieID of user and the UID of correspondence, and be sent to user data acquisition module 601.User data acquisition module 601 can according to the UID of the cookieID of user and correspondence, obtains the user behavior data of user in setting cycle and user social contact data and exports.
Advertising inquiry module 802 for class of subscriber is divided the user that module 603 exports category division result as search keyword, find out and the ad content searched for keyword and match.
After the ad content that request processing module 803 finds out for insertion of advertising enquiry module 802 in the advertisement played column of webpage of asking in request of access, the webpage of request of access being asked returns to user.
In the embodiment of the present invention, the concrete function realization of each module in the category division system of user, unit, subelement is carried out based on Social behaviors, can with reference to the detailed process of step S101-103, step S301-S305 and the step S401-S405 in the above-mentioned category classification method carrying out user based on Social behaviors; The concrete function of each module in advertisement delivery system realizes, and with reference to the step S201-204 in above-mentioned advertisement placement method, can be not described in detail in this.
In technical scheme of the present invention, according to the user behavior data of user on website, user social contact data under social platform, extract user behavior attributive character and user social contact attributive character respectively; Then, the category division of user is carried out according to the user behavior attributive character extracted and user social contact attributive character.Compare existing only according to the anti-interest pushing away user of the information of user to access pages, technical scheme provided by the invention, consider user behavior attributive character and user social contact attributive character to carry out class of subscriber division, the feature rich degree of category division can be increased, and improve class of subscriber division accuracy, also just can improve crowd's positional accuracy; Based on higher crowd's positional accuracy, the specific aim of the audient of advertisement, promotional videos etc. can be improved, throw in validity.
The above is only the preferred embodiment of the present invention; it should be pointed out that for those skilled in the art, under the premise without departing from the principles of the invention; can also make some improvements and modifications, these improvements and modifications also should be considered as protection scope of the present invention.
Claims (10)
1. carry out a user's category classification method based on Social behaviors, it is characterized in that, comprising:
Obtain the user behavior data of user to be sorted in setting cycle, and user social contact data;
From described user behavior data, extract user behavior attributive character, from described user social contact data, extract user social contact attributive character;
According to the user behavior attributive character extracted and user social contact attributive character, described user is divided into respective classes.
2. the method for claim 1, is characterized in that, the described user behavior attributive character according to extraction and user social contact attributive character, be divided into respective classes by described user, specifically comprise:
By the user behavior attributive character of described user, mate with the user behavior attributive character recorded in the behavioural characteristic rule base built in advance, judge the behavior group belonging to described user according to matching result; Wherein, for Mei Zhong behavior group in described behavioural characteristic rule base, have recorded the behavior user behavior attributive character that has of group;
By the user social contact attributive character of described user, mate with the user social contact attributive character recorded in the social characteristics rule base built in advance, judge the social group belonging to described user according to matching result; Wherein, for often kind of social group in described social characteristics rule base, have recorded the user social contact attributive character that this social group has;
Find out group common between behavior group belonging to described user and social group, and described user is divided in the common group found out.
3. the method for claim 1, is characterized in that, the described user behavior attributive character according to extraction and user social contact attributive character, be divided into respective classes by described user, specifically comprise:
The user behavior attributive character of the described user extracted and user social contact attributive character are merged and gathers the user characteristics obtaining described user;
The user characteristics of the described user obtained is input in the category division model for category division; By described category division model, described user is divided into respective classes.
4. method as claimed in claim 3, it is characterized in that, described category division model is training in advance:
For each training user selected from the user of website, after the user behavior attributive character of adding up this training user and user social contact attributive character, merge and gather the user characteristics obtaining this training user; And determine the behavior group belonging to this training user according to the user behavior attributive character of this training user, the social group belonging to this training user is determined according to the user social contact attributive character of this training user, and using group common between the behavior group belonging to this training user and social group as group belonging to this training user;
Group, user characteristics belonging to each training user, as training data, are used described training data to use preset many labels multi-classification algorithm to carry out model training, obtain category division model.
5. method as claimed in claim 4, is characterized in that, described obtain category division model before, also comprise:
For each test subscriber selected from the user of website, after the user behavior attributive character of adding up this test subscriber and user social contact attributive character, merge and gather the user characteristics obtaining this test subscriber; And determine the behavior group belonging to this test subscriber according to the user behavior attributive character of this test subscriber, the social group belonging to this test subscriber is determined according to the user social contact attributive character of this test subscriber, and using group common between the behavior group belonging to this test subscriber and social group as group belonging to this test subscriber;
Using group, user characteristics belonging to each test subscriber as test data; And
When using described training data to use preset many labels multi-classification algorithm to carry out model training, also using described test data to use preset many labels multi-classification algorithm to carry out model training, obtaining described category division model.
6. an advertisement placement method, is characterized in that, comprising:
After website receives the request of access of user, the user identity of the identification number cookieID and correspondence that obtain described user proves UID;
According to cookieID and UID of described user, adopt as arbitrary in claim 1-5 as described in method, obtain the category division result of described user;
Using the category division result that obtains as search keyword, find out the ad content matched with described search keyword;
The ad content found out is play in the advertisement played column of the webpage returned to described user for described request of access.
7. carry out a user's category division system based on Social behaviors, it is characterized in that, comprising:
User data acquisition module, for obtaining the user behavior data of user to be sorted in setting cycle and user social contact data and exporting;
Attributive character abstraction module, for extracting user behavior attributive character in the user behavior data from described user data acquisition module output, extracts user social contact attributive character from the user social contact data that described user data acquisition module exports;
Class of subscriber divides module, for the user behavior attributive character that extracts according to described attributive character abstraction module and user social contact attributive character, described user is divided into respective classes.
8. system as claimed in claim 7, is characterized in that, described class of subscriber divides module and specifically comprises: first user category division unit or the second class of subscriber division unit; Wherein,
Described first user category division unit specifically comprises:
Behavioural characteristic coupling subelement, for the user behavior attributive character that described attributive character abstraction module is extracted, mate with the user behavior attributive character recorded in the behavioural characteristic rule base built in advance, judge the behavior group belonging to described user according to matching result; Wherein, for Mei Zhong behavior group in described behavioural characteristic rule base, have recorded the behavior user behavior attributive character that has of group;
Social characteristics coupling subelement, for the user social contact attributive character that described attributive character abstraction module is extracted, mate with the user social contact attributive character recorded in the social characteristics rule base built in advance, judge the social group belonging to described user according to matching result; Wherein, for often kind of social group in described social characteristics rule base, have recorded the user social contact attributive character that this social group has;
Category division subelement, for receive described behavioural characteristic matching unit export user belonging to behavior group and described social characteristics matching unit export user belonging to social group; Find out group common between behavior group belonging to described user and social group, and described user is divided in the common group found out; And
Described second class of subscriber division unit is used for user behavior attributive character and the user social contact attributive character of the described user extracted by described attributive character abstraction module, merges and gathers the user characteristics obtaining described user; And the user characteristics of the described user obtained is input in the category division model for category division; By described category division model, described user is divided into respective classes.
9. system as claimed in claim 6, is characterized in that, also comprise:
Partitioning model training module, for for each training user selected from the user of website, after the user behavior attributive character of adding up this training user and user social contact attributive character, merges and gathers the user characteristics obtaining this training user; And determine the behavior group belonging to this training user according to the user behavior attributive character of this training user, the social group belonging to this training user is determined according to the user social contact attributive character of this training user, and using group common between the behavior group belonging to this training user and social group as group belonging to this training user; Group, user characteristics belonging to each training user, as training data, are used described training data to use preset many labels multi-classification algorithm to carry out model training, obtain category division model.
10. an advertisement delivery system, is characterized in that, comprising:
User data acquisition module as described in as arbitrary in claim 7-9, attributive character abstraction module, class of subscriber divide module;
Web-page requests receiver module, for receive user request of access after, obtain the cookieID of user and the UID of correspondence, and be sent to described user data acquisition module;
Advertising inquiry module, for described class of subscriber being divided the category division result of the user that module exports as search keyword, finds out the ad content matched with described search keyword;
Request processing module, after inserting ad content that described advertising inquiry module searches goes out in the advertisement played column of webpage of asking in described request of access, returns described webpage to described user;
Described user data acquisition module specifically for according to the cookieID of user and the UID of correspondence, obtains the user behavior data of user in setting cycle and user social contact data and exports.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410492126.5A CN104298719B (en) | 2014-09-23 | 2014-09-23 | Category division, advertisement placement method and the system of user is carried out based on Social behaviors |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410492126.5A CN104298719B (en) | 2014-09-23 | 2014-09-23 | Category division, advertisement placement method and the system of user is carried out based on Social behaviors |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104298719A true CN104298719A (en) | 2015-01-21 |
| CN104298719B CN104298719B (en) | 2018-02-27 |
Family
ID=52318444
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410492126.5A Active CN104298719B (en) | 2014-09-23 | 2014-09-23 | Category division, advertisement placement method and the system of user is carried out based on Social behaviors |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104298719B (en) |
Cited By (31)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104683122A (en) * | 2015-02-12 | 2015-06-03 | 北京集奥聚合科技有限公司 | Multi-screen-linkage-based information transmitting method and system |
| CN104951544A (en) * | 2015-06-19 | 2015-09-30 | 百度在线网络技术（北京）有限公司 | User data processing method and system and method and system for providing user data |
| CN104992342A (en) * | 2015-05-11 | 2015-10-21 | 腾讯科技（北京）有限公司 | Method for determining the effectiveness of popularizing information release, monitoring server, and terminal |
| CN105447376A (en) * | 2015-10-30 | 2016-03-30 | 广州市汇助惠电子商务有限公司 | User management system |
| CN105516928A (en) * | 2016-01-15 | 2016-04-20 | 中国联合网络通信有限公司广东省分公司 | Position recommending method and system based on position crowd characteristics |
| WO2016155260A1 (en) * | 2015-04-02 | 2016-10-06 | 百度在线网络技术（北京）有限公司 | Information delivery method, device and system |
| CN106204083A (en) * | 2015-04-30 | 2016-12-07 | 中国移动通信集团山东有限公司 | A kind of targeted customer's sorting technique, Apparatus and system |
| CN106294508A (en) * | 2015-06-10 | 2017-01-04 | 深圳市腾讯计算机系统有限公司 | A kind of brush amount tool detection method and device |
| CN106487825A (en) * | 2015-08-25 | 2017-03-08 | 阿里巴巴集团控股有限公司 | information correlation method and device |
| CN106600323A (en) * | 2016-12-13 | 2017-04-26 | 合肥华耀广告传媒有限公司 | Advertisement sharing system based on social network |
| CN106780064A (en) * | 2016-12-02 | 2017-05-31 | 腾讯科技（深圳）有限公司 | A kind of region partitioning method, device and the network equipment |
| CN106817384A (en) * | 2015-11-27 | 2017-06-09 | 亿阳信通股份有限公司 | A kind of analysis method and system that behavior is accessed based on user's telecommunications |
| CN107205011A (en) * | 2016-03-17 | 2017-09-26 | 阿里巴巴集团控股有限公司 | Business datum method for pushing and its system |
| CN107330001A (en) * | 2017-06-09 | 2017-11-07 | 国政通科技股份有限公司 | The creation method and system of a kind of diversification label |
| CN107392648A (en) * | 2017-06-29 | 2017-11-24 | 上海精数信息科技有限公司 | Assess the method and device of directory audient distribution |
| CN107451247A (en) * | 2017-07-28 | 2017-12-08 | 北京小米移动软件有限公司 | user identification method and device |
| CN108182255A (en) * | 2017-12-29 | 2018-06-19 | 重庆金融资产交易所有限责任公司 | Property right project information recommends method, apparatus, storage medium and computer equipment |
| CN108280458A (en) * | 2017-01-05 | 2018-07-13 | 腾讯科技（深圳）有限公司 | Group relation kind identification method and device |
| CN109033118A (en) * | 2018-05-23 | 2018-12-18 | 国政通科技股份有限公司 | A kind of object-based dynamic data judgment method and device |
| CN109345075A (en) * | 2018-08-31 | 2019-02-15 | 深圳市轱辘汽车维修技术有限公司 | A kind of professional person, which authenticates, investigates method, apparatus and terminal device |
| CN109409943A (en) * | 2018-10-11 | 2019-03-01 | 温州你创我帮网络科技有限公司 | A kind of advertisement promotion system Internet-based |
| TWI670662B (en) * | 2017-11-09 | 2019-09-01 | 財團法人資訊工業策進會 | Inference system for data relation, method and system for generating marketing targets |
| CN110197386A (en) * | 2018-04-12 | 2019-09-03 | 腾讯科技（深圳）有限公司 | Media resource method for pushing and device, storage medium and electronic device |
| CN110917628A (en) * | 2018-09-20 | 2020-03-27 | 北京默契破冰科技有限公司 | Method, apparatus, and computer storage medium for determining user grouping |
| CN110917629A (en) * | 2018-09-20 | 2020-03-27 | 北京默契破冰科技有限公司 | Method, apparatus, and computer storage medium for managing user behavior |
| CN111353001A (en) * | 2018-12-24 | 2020-06-30 | 杭州海康威视数字技术股份有限公司 | Method and device for classifying users |
| CN112435067A (en) * | 2020-11-30 | 2021-03-02 | 翼果（深圳）科技有限公司 | Intelligent advertisement putting method and system for cross-e-commerce platform and social platform |
| CN113630336A (en) * | 2021-07-19 | 2021-11-09 | 上海德衡数据科技有限公司 | Data distribution method and system based on optical interconnection |
| CN113672777A (en) * | 2021-08-30 | 2021-11-19 | 上海飞旗网络技术股份有限公司 | User intention exploration method and system based on traffic correlation analysis |
| TWI757854B (en) * | 2020-08-28 | 2022-03-11 | 中國信託商業銀行股份有限公司 | Business recommendation system and method |
| US11843651B2 (en) | 2019-04-03 | 2023-12-12 | Huawei Technologies Co., Ltd. | Personalized recommendation method and system, and terminal device |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101685521A (en) * | 2008-09-23 | 2010-03-31 | 北京搜狗科技发展有限公司 | Method for showing advertisements in webpage and system |
| US8170971B1 (en) * | 2011-09-28 | 2012-05-01 | Ava, Inc. | Systems and methods for providing recommendations based on collaborative and/or content-based nodal interrelationships |
| CN102663026A (en) * | 2012-03-22 | 2012-09-12 | 浙江盘石信息技术有限公司 | Implementation method for directionally running internet advertisements |
| CN103366020A (en) * | 2013-08-06 | 2013-10-23 | 刘临 | System and method for analyzing user behaviors |
| CN103458042A (en) * | 2013-09-10 | 2013-12-18 | 上海交通大学 | Microblog advertisement user detection method |
| CN103778555A (en) * | 2014-01-21 | 2014-05-07 | 北京集奥聚合科技有限公司 | User attribute mining method and system based on user tags |
-
2014
- 2014-09-23 CN CN201410492126.5A patent/CN104298719B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101685521A (en) * | 2008-09-23 | 2010-03-31 | 北京搜狗科技发展有限公司 | Method for showing advertisements in webpage and system |
| US8170971B1 (en) * | 2011-09-28 | 2012-05-01 | Ava, Inc. | Systems and methods for providing recommendations based on collaborative and/or content-based nodal interrelationships |
| CN102663026A (en) * | 2012-03-22 | 2012-09-12 | 浙江盘石信息技术有限公司 | Implementation method for directionally running internet advertisements |
| CN103366020A (en) * | 2013-08-06 | 2013-10-23 | 刘临 | System and method for analyzing user behaviors |
| CN103458042A (en) * | 2013-09-10 | 2013-12-18 | 上海交通大学 | Microblog advertisement user detection method |
| CN103778555A (en) * | 2014-01-21 | 2014-05-07 | 北京集奥聚合科技有限公司 | User attribute mining method and system based on user tags |
Cited By (44)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104683122A (en) * | 2015-02-12 | 2015-06-03 | 北京集奥聚合科技有限公司 | Multi-screen-linkage-based information transmitting method and system |
| WO2016155260A1 (en) * | 2015-04-02 | 2016-10-06 | 百度在线网络技术（北京）有限公司 | Information delivery method, device and system |
| CN106204083A (en) * | 2015-04-30 | 2016-12-07 | 中国移动通信集团山东有限公司 | A kind of targeted customer's sorting technique, Apparatus and system |
| CN104992342B (en) * | 2015-05-11 | 2019-08-13 | 腾讯科技（北京）有限公司 | Promotion message delivery effectiveness determines method, monitoring server and terminal |
| CN104992342A (en) * | 2015-05-11 | 2015-10-21 | 腾讯科技（北京）有限公司 | Method for determining the effectiveness of popularizing information release, monitoring server, and terminal |
| US10719847B2 (en) | 2015-05-11 | 2020-07-21 | Tencent Technology (Shenzhen) Company Limited | Method for determining validity of delivering of promotion information, monitoring server and terminal |
| CN106294508B (en) * | 2015-06-10 | 2020-02-11 | 深圳市腾讯计算机系统有限公司 | Brushing amount tool detection method and device |
| CN106294508A (en) * | 2015-06-10 | 2017-01-04 | 深圳市腾讯计算机系统有限公司 | A kind of brush amount tool detection method and device |
| CN104951544A (en) * | 2015-06-19 | 2015-09-30 | 百度在线网络技术（北京）有限公司 | User data processing method and system and method and system for providing user data |
| CN106487825A (en) * | 2015-08-25 | 2017-03-08 | 阿里巴巴集团控股有限公司 | information correlation method and device |
| CN106487825B (en) * | 2015-08-25 | 2019-12-24 | 阿里巴巴集团控股有限公司 | Information association method and device |
| CN105447376B (en) * | 2015-10-30 | 2018-08-03 | 广州市汇助惠电子商务有限公司 | A kind of Subscriber Management System |
| CN105447376A (en) * | 2015-10-30 | 2016-03-30 | 广州市汇助惠电子商务有限公司 | User management system |
| CN106817384A (en) * | 2015-11-27 | 2017-06-09 | 亿阳信通股份有限公司 | A kind of analysis method and system that behavior is accessed based on user's telecommunications |
| CN105516928A (en) * | 2016-01-15 | 2016-04-20 | 中国联合网络通信有限公司广东省分公司 | Position recommending method and system based on position crowd characteristics |
| CN107205011A (en) * | 2016-03-17 | 2017-09-26 | 阿里巴巴集团控股有限公司 | Business datum method for pushing and its system |
| CN106780064B (en) * | 2016-12-02 | 2021-01-05 | 腾讯科技（深圳）有限公司 | Region division method, device and network equipment |
| CN106780064A (en) * | 2016-12-02 | 2017-05-31 | 腾讯科技（深圳）有限公司 | A kind of region partitioning method, device and the network equipment |
| CN106600323A (en) * | 2016-12-13 | 2017-04-26 | 合肥华耀广告传媒有限公司 | Advertisement sharing system based on social network |
| CN108280458A (en) * | 2017-01-05 | 2018-07-13 | 腾讯科技（深圳）有限公司 | Group relation kind identification method and device |
| CN108280458B (en) * | 2017-01-05 | 2022-01-14 | 腾讯科技（深圳）有限公司 | Group relation type identification method and device |
| CN107330001A (en) * | 2017-06-09 | 2017-11-07 | 国政通科技股份有限公司 | The creation method and system of a kind of diversification label |
| CN107392648A (en) * | 2017-06-29 | 2017-11-24 | 上海精数信息科技有限公司 | Assess the method and device of directory audient distribution |
| CN107451247B (en) * | 2017-07-28 | 2021-03-30 | 北京小米移动软件有限公司 | User identification method and device |
| CN107451247A (en) * | 2017-07-28 | 2017-12-08 | 北京小米移动软件有限公司 | user identification method and device |
| TWI670662B (en) * | 2017-11-09 | 2019-09-01 | 財團法人資訊工業策進會 | Inference system for data relation, method and system for generating marketing targets |
| CN108182255A (en) * | 2017-12-29 | 2018-06-19 | 重庆金融资产交易所有限责任公司 | Property right project information recommends method, apparatus, storage medium and computer equipment |
| CN108182255B (en) * | 2017-12-29 | 2020-07-28 | 重庆金融资产交易所有限责任公司 | Title item information recommendation method and device, storage medium and computer equipment |
| CN110197386A (en) * | 2018-04-12 | 2019-09-03 | 腾讯科技（深圳）有限公司 | Media resource method for pushing and device, storage medium and electronic device |
| CN110197386B (en) * | 2018-04-12 | 2022-02-08 | 腾讯科技（深圳）有限公司 | Media resource pushing method and device, storage medium and electronic device |
| CN109033118A (en) * | 2018-05-23 | 2018-12-18 | 国政通科技股份有限公司 | A kind of object-based dynamic data judgment method and device |
| CN109033118B (en) * | 2018-05-23 | 2021-06-29 | 国政通科技股份有限公司 | Dynamic data judgment method and device based on object |
| CN109345075A (en) * | 2018-08-31 | 2019-02-15 | 深圳市轱辘汽车维修技术有限公司 | A kind of professional person, which authenticates, investigates method, apparatus and terminal device |
| CN110917629A (en) * | 2018-09-20 | 2020-03-27 | 北京默契破冰科技有限公司 | Method, apparatus, and computer storage medium for managing user behavior |
| CN110917628A (en) * | 2018-09-20 | 2020-03-27 | 北京默契破冰科技有限公司 | Method, apparatus, and computer storage medium for determining user grouping |
| CN109409943A (en) * | 2018-10-11 | 2019-03-01 | 温州你创我帮网络科技有限公司 | A kind of advertisement promotion system Internet-based |
| CN111353001A (en) * | 2018-12-24 | 2020-06-30 | 杭州海康威视数字技术股份有限公司 | Method and device for classifying users |
| CN111353001B (en) * | 2018-12-24 | 2023-08-18 | 杭州海康威视数字技术股份有限公司 | Method and device for classifying users |
| US11843651B2 (en) | 2019-04-03 | 2023-12-12 | Huawei Technologies Co., Ltd. | Personalized recommendation method and system, and terminal device |
| TWI757854B (en) * | 2020-08-28 | 2022-03-11 | 中國信託商業銀行股份有限公司 | Business recommendation system and method |
| CN112435067A (en) * | 2020-11-30 | 2021-03-02 | 翼果（深圳）科技有限公司 | Intelligent advertisement putting method and system for cross-e-commerce platform and social platform |
| CN113630336A (en) * | 2021-07-19 | 2021-11-09 | 上海德衡数据科技有限公司 | Data distribution method and system based on optical interconnection |
| CN113672777A (en) * | 2021-08-30 | 2021-11-19 | 上海飞旗网络技术股份有限公司 | User intention exploration method and system based on traffic correlation analysis |
| CN113672777B (en) * | 2021-08-30 | 2023-09-08 | 上海飞旗网络技术股份有限公司 | User intention exploration method and system based on flow correlation analysis |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104298719B (en) | 2018-02-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104298719A (en) | Method and system for conducting user category classification and advertisement putting based on social behavior | |
| US10997265B1 (en) | Selecting a template for a content item | |
| CN103473273B (en) | Information search method, device and server | |
| CN108197330B (en) | Data digging method and device based on social platform | |
| US8676875B1 (en) | Social media measurement | |
| US20140280549A1 (en) | Method and System for Efficient Matching of User Profiles with Audience Segments | |
| US20160188725A1 (en) | Method and System for Enhanced Content Recommendation | |
| US20160180402A1 (en) | Method for recommending products based on a user profile derived from metadata of multimedia content | |
| CN101577866A (en) | User classification method, advertisement release method and device | |
| US20170039205A1 (en) | Dynamic radius threshold selection | |
| CN102982153A (en) | Information retrieval method and device | |
| CN101398926A (en) | Advertisement bidding and playing method and system based on directional delivery | |
| CN104462336A (en) | Information pushing method and device | |
| CN104574130A (en) | Accurate advertisement injecting method and system based on customer resource library | |
| CN106033415A (en) | A text content recommendation method and device | |
| US20130066708A1 (en) | Online advertising system and a method of operating the same | |
| US20210097045A1 (en) | Object identifier index | |
| CN105916032A (en) | Video recommendation method and video recommendation terminal equipment | |
| US20130006760A1 (en) | Systems and methods for presenting comparative advertising | |
| CN102999521B (en) | A kind of method and device identifying search need | |
| CN103745380A (en) | Advertisement delivery method and apparatus | |
| CN104281641A (en) | Method for enriching a multimedia content, and corresponding device | |
| CN103955480A (en) | Method and equipment for determining target object information corresponding to user | |
| CN106844792B (en) | Method and system for realizing advertisement of primary information designated audience of social relationship | |
| US20140214541A1 (en) | Method and system for user-controlled rendering of mobile advertisements |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20230417 Address after: Room 501-502, 5/F, Sina Headquarters Scientific Research Building, Block N-1 and N-2, Zhongguancun Software Park, Dongbei Wangxi Road, Haidian District, Beijing, 100193 Patentee after: Sina Technology (China) Co.,Ltd. Address before: 100080, International Building, No. 58 West Fourth Ring Road, Haidian District, Beijing, 20 floor Patentee before: Sina.com Technology (China) Co.,Ltd. |
|
| TR01 | Transfer of patent right |