CN103489117B - Method and system for information releasing - Google Patents
Method and system for information releasing Download PDFInfo
- Publication number
- CN103489117B CN103489117B CN201210192864.9A CN201210192864A CN103489117B CN 103489117 B CN103489117 B CN 103489117B CN 201210192864 A CN201210192864 A CN 201210192864A CN 103489117 B CN103489117 B CN 103489117B
- Authority
- CN
- China
- Prior art keywords
- user
- information
- user behavior
- time
- feature
- 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.)
- Active
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
A method for information releasing is characterized by comprising the following steps that user behavior characteristics and information characteristics corresponding to the user behavior characteristics of a user are recorded; weights of different user behavior characteristics are obtained; time of the user behavior characteristics is obtained, time weights are determined according to time, the closer the time of user behaviors is, the higher the time weights are, and the farther the time is, the lower the time weights are; the weighted value of the same-kind user behavior characteristics of the same user at different periods of time is determined according to the weights of the different user behavior characteristics and the time weights; according to information characteristics of multi-preset-information to be released, corresponding relations of information characteristics of the user behavior characteristics and the weighted values of the user behavior characteristics, the probability value of the clicked preset information containing the information characteristics is calculated; the preset information is pushed to the user according to the probability value. The invention further provides a corresponding information releasing system. The probability value figured out through the method embodies current interestingness of the user, and thus more accurate information releasing is achieved.
Description
Technical field
The present invention relates to field of computer technology, particularly relate to information distribution method and the system of internet.
Background technology
Along with the fast development of internet, netizen's quantity rapidly expands, and the web advertisement is more and more subject to liking of advertiser, and its injected volume has exceeded the conventional ads such as newspaper, TV.Current charging mode mainly contains CPM(by exposure charge), CPD(daily charges), CPC(is by clicking charge), CPA(is by effect charge), CPS(deducts a percentage by sales volume) etc., for the medium and small advertiser of major part, the pattern that CPC, CPA etc. pay by effect is more and more welcome.Promote the clicking rate of advertisement, the satisfaction of advertiser can not only be promoted, go back and can directly promote advertising platform income.
Current advertisement putting rule has:
1) to be mated with advertisement position specification by advertisement and content of pages mates, decide advertisement putting.This method can solve mating of advertisement and advertisement position, but to different user, display advertisement is identical, personalized disappearance, limited to clicking rate lifting degree.
2) the population-based characteristic row orientation couplings such as age of user, sex, region are used.The method introduces certain customers' feature, but this part feature is static nature substantially, and the long period is constant, also just causes user to show same advertisement for a long time.And population-based feature has certain deviation with click interest wish, low to the discrimination of advertisement.Such as, not all women's any time all likes best clicks women's dress advertisement, and some moment of some women also may like game advertisement.
A lot of ad system employs gathering of above method, but lacks the expansion to many time varying characteristics, and user profile is lacked to some extent, can not calculate the matching degree of user and advertisement more accurately.
Summary of the invention
Based on this, provide the more accurate information distribution method of one and system.
A kind of information distribution method, comprises the steps:
The user behavior feature of counting user and the information characteristics of correspondence;
Obtain the weight of different user behavioural characteristic;
Obtain the time of described user behavior feature, determine time weighting according to the described time, the time of described user behavior is nearer, and described time weighting is larger, and the time is far away, and described time weighting is less;
According to weight and the described time weighting of described different user behavioural characteristic, determine the weighted value of the user behavior feature of the same race of the different time of same user;
According to multiple for the information characteristics of presupposed information of input and the corresponding relation of the information characteristics of described user behavior feature, and the weighted value of user behavior feature described in each, calculate the probable value that each presupposed information comprising described information characteristics is clicked;
According to described probable value, described presupposed information is pushed to described user.
In one of them embodiment, described user behavior feature comprise following at least one: browse, collect, place an order, strike a bargain, APP information, the login of game, the registration of game, game download, to post containing the microblogging of certain keyword or to forward.
In one of them embodiment, the user behavior feature of described counting user and the information characteristics of correspondence comprise: the user ID obtaining described user, obtain and add up the user behavior feature of described user at heterogeneous networks platform and the information characteristics of correspondence according to described user ID.
In one of them embodiment, the user behavior feature of described counting user and the information characteristics of correspondence carry out in a preset time period, and different user behavior features adopts different preset time period to add up.
In one of them embodiment, described preset time period length is based on the available time of described user behavior, and available time is longer, and preset time period is longer.
In one of them embodiment, this information that the weight of described user behavior feature shows based on user behavior is to the importance of described user, and importance is higher, and weight is larger.
In one of them embodiment, describedly according to described probable value, described presupposed information is pushed to described user and comprises: at least one presupposed information before the highest for probable value is pushed to described user.
A kind of information jettison system, it comprises: feature acquisition module, Weight Acquisition module, time weighting acquisition module, assignment module, probability evaluation entity and info push module,
Described feature acquisition module is used for the user behavior feature of counting user and the information characteristics of correspondence;
Described Weight Acquisition module is for obtaining the weight of different user behavioural characteristic;
Described time weighting acquisition module, for obtaining the time of described user behavior feature, determines time weighting according to the described time, and the time of described user behavior is nearer, and described time weighting is larger, and the time is far away, and described time weighting is less;
Described assignment module is used for, according to the weight of described different user behavioural characteristic and described time weighting, determining the weighted value of the user behavior feature of the same race of the different time of same user;
Described probability evaluation entity is used for multiple for the information characteristics of presupposed information of input and the corresponding relation of the information characteristics of described user behavior feature according to what preset, and the weighted value of user behavior feature described in each, calculate each and comprise the clicked probable value of described information characteristics presupposed information;
Described info push module is used for, according to described probable value, default presupposed information is pushed to described user.
In one of them embodiment, described user behavior feature comprise browse, collect, place an order, strike a bargain, APP information, the login of game, the registration of game, game download, to post containing the microblogging of certain keyword or at least one in forwarding.
In one of them embodiment, described feature acquisition module is used for the user ID first obtaining described user, then obtains according to described user ID and add up the user behavior feature of described user at heterogeneous networks platform and the information characteristics of correspondence.
In one of them embodiment, the described user behavior feature of feature acquisition module counting user and the information characteristics of correspondence carry out in a preset time period, and different user behavior features adopts different preset time period to add up.
In one of them embodiment, described preset time period length is based on the available time of described user behavior, and available time is longer, and preset time period is longer.
In one of them embodiment, this information that the different weights of described default different user behavioural characteristic show based on user behavior is to the importance of described user, and importance is higher, and weight is larger.
In one of them embodiment, describedly according to described probable value, described presupposed information is pushed to described user and comprises: according to described probable value extract probable value the highest before at least one presupposed information be pushed to described user.
Above-mentioned information distribution method and system, by different user behavior features, different weights is set, and successively different weighted values is arranged to the generation time of same user behavior feature, the probable value calculated more can embody the current interest-degree of user, thus the information more accurately that realizes is thrown in.
Accompanying drawing explanation
Fig. 1 is the flow chart of steps of the information distribution method of an embodiment;
Fig. 2 is the functional block diagram of the information jettison system of an embodiment;
Fig. 3 describes from macroscopic perspective the prediction flow process Organization Chart clicking probability;
Fig. 4 gathers the integration process flow diagram of user behavior feature in layer for the feature shown in Fig. 3.
Specific embodiment
The object of this programme is: one or several presupposed informations calculating user interest degree higher (clicking probability larger) from the multiple presupposed informations for throwing in, and shows user.Realize this function, can the feature of reference comprise:
1, information characteristics, such as classification, price, title etc.;
2, information throws in window feature, the such as classification etc. of size, the affiliated page;
3, user's static nature, such as age, region, sex, educational background etc.;
4, user behavior feature, such as browse, collect, place an order, strike a bargain, APP information, the login of game, the registration of game, game download, to post containing the microblogging of certain keyword or forwarding etc.
First three kind feature can be directly used in filters the multiple presupposed informations for throwing in, and obtains the presupposed information of candidate.User behavior feature, clicks the probability size of each presupposed information for user in the presupposed information of calculated candidate.
As shown in Figure 1, it is the flow chart of steps of the information distribution method of an embodiment, comprises the steps:
Step S101, the user behavior feature of counting user and the information characteristics of correspondence.
Described user behavior feature comprises " browse, collect, place an order, strike a bargain, APP information, the login of game, the registration of game, game download, to post containing the microblogging of certain keyword or to forward " at least one.In the present embodiment, with the feature being preferably commodity for information characteristics corresponding in behavioural characteristic, such as " commodity classification ", as " shoes ", " skirt ", " mobile phone " etc., also can be the classification specific to certain brand, such as " Samsung mobile phone ".User behavior integrate features information characteristics can enumerate example as browsed shoes, and collection skirt, place an order mobile phone etc.
In other embodiments, the user behavior feature of described counting user and the information characteristics of correspondence carry out in a preset time period, and different user behavior features adopts different preset time period to add up.Described preset time period length is based on the available time of described user behavior, and available time is longer, and preset time period is longer.The principle that in one embodiment, preset time period length is followed " browse, collect, place an order, strike a bargain " reduces successively.As: counting user " browsing " information of nearly 15 days, " collection " information of nearly 10 days, " placing an order " information of nearly 5 days, and " conclusion of the business " information of nearly 2 days.
Because the behavior that browsing is a kind of longer-term, consciousness property is not strong, needs the even monthly cycle accumulation in a lot of sky to reflect the Long-term Interest of user.And to place an order be strong, the Impulsive high behavior of a kind of property realized, if the time has been grown, user has completed conclusion of the business, may all can not be interested in this type of information again recently.So according to the statistics that the different time cycle carries out different user behavioural characteristic, the current interest of user can be indicated more accurately.
In the present embodiment, the user behavior feature of described counting user and the information characteristics of correspondence comprise: the user ID obtaining described user, obtain and add up the user behavior feature of described user at heterogeneous networks platform and the information characteristics of correspondence according to described user ID.As by the action message of user on each platform of having connected such as cookie or alliance website, so just can accomplish more transfer learning, by the performance of user on other platforms, reflect user's possible interest on the platform.The user behavior feature of different platform can arrange different weights.
The life cycle of information recommendation is often very short, and for advertisement as wish impression information, it on average only has several days, and some advertisements even only have 1,2 day.Such dependence advertisement self is thrown in history and is predicted click probability, and because its behavior property is few, cover customer volume low, accuracy rate is limited.Experiment proves, the behavioural characteristic of user has stronger relevance with ad click usually, so our user characteristics of just other platforms of consideration introducing predicts the click wish of user to advertisement.Such as, user has just collected a women's dress on electric business's platform, and represent it and be interested in women's dress, and when this user browses portal website's platform, throw in the relevant advertisement of women's dress to this user, it clicks wish will be stronger.This related platform is more, and the user characteristics that we can catch will be more, and the click wish of prediction user to which advertisement will be more accurate.
Step S102, obtains the weight of different user behavioural characteristic.
This information that the weight of described user behavior feature shows based on user behavior is to the importance of described user, and importance is higher, and weight is larger.As " browsing " once, count 5 points, " collection " once counts 10 points, and " placing an order " once counts 50 points, and " conclusion of the business " once counts 80 points.In the present embodiment, the weighted value of " browse, collect, place an order, strike a bargain " from low to high.
Step S103, obtain the time of described user behavior feature, determine time weighting according to the described time, the time of described user behavior is nearer, and described time weighting is larger, and the time is far away, and described time weighting is less.
Because nearer browsing more can reflect the hobby that user is nearest, should give and higher weight, and behavior far away, should give and lower weight.For " placing an order ", the conclusion of the business of nearest first day, each meter 50 points, the conclusion of the business of nearest second day, each meter 40 points, the conclusion of the business of nearest the 3rd day, each score 30 points ...Namely assignment decayed from the near to the remote according to the time, can be realized by time attenuation function.
Step S104, according to weight and the described time weighting of described different user behavioural characteristic, determines the weighted value of the user behavior feature of the same race of the different time of same user.
Step S105, according to multiple for the information characteristics of presupposed information of input and the corresponding relation of the information characteristics of described user behavior feature, and the weighted value of user behavior feature described in each, calculate each and comprise the clicked probable value of described information characteristics presupposed information.
In one embodiment, step S105 also in conjunction with user's static nature (age, sex etc.), information characteristics (classification, price etc.), can utilize classification or regression model, calculates the click probability of user to each presupposed information.
Step S106, is pushed to described user according to described probable value by described presupposed information.
Describedly according to described probable value, described presupposed information is pushed to described user and comprises: at least one presupposed information before the highest for probable value is pushed to described user.
Above-mentioned information distribution method arranges different weights by different user behavior features, and successively different weighted values is arranged to the generation time of same user behavior feature, the probable value calculated more can embody the current interest-degree of user, thus the information more accurately that realizes is thrown in.
In other embodiments, described information distribution method can also comprise: by information characteristics, and information throws in window feature, and the step that at least one in user's static nature is screened multiple presupposed information.
As shown in Figure 2, it is the functional block diagram of the information jettison system 20 of an embodiment, comprising: feature acquisition module 202, Weight Acquisition module 203, time weighting acquisition module 204, assignment module 205, probability evaluation entity 206 and info push module 207.
The user behavior feature of feature acquisition module 202 for counting user and the information characteristics of correspondence.
Weight Acquisition module 203 is for obtaining the weight of different user behavioural characteristic.
Time weighting acquisition module 204, for obtaining the time of described user behavior feature, determines time weighting according to the described time, and the time of described user behavior is nearer, and described time weighting is larger, and the time is far away, and described time weighting is less.
Assignment module 205, for according to the weight of described different user behavioural characteristic and described time weighting, determines the weighted value of the user behavior feature of the same race of the different time of same user.
Probability evaluation entity 206 is for multiple for the information characteristics of presupposed information of input and the corresponding relation of the information characteristics of described user behavior feature according to what preset, and the weighted value of user behavior feature described in each, calculate each and comprise the clicked probable value of described information characteristics presupposed information.
Info push module 207 is for being pushed to described user according to described probable value by default presupposed information.
In one embodiment, the information characteristics of described default presupposed information and correspondence thereof is stored in a memory module.
In one embodiment, describedly according to described probable value, described presupposed information is pushed to described user and comprises: according to described probable value extract from described memory module probable value the highest before at least one presupposed information be pushed to described user.
In one embodiment, feature acquisition module 202, for obtaining the user ID of described user, obtains according to described user ID and adds up the user behavior feature of described user at heterogeneous networks platform and the information characteristics of correspondence.
Described user behavior feature comprises " browse, collect, place an order, strike a bargain, APP information, the login of game, the registration of game, game download, to post containing the microblogging of certain keyword or to forward " at least one.
The described user behavior feature of feature acquisition module 202 counting user and the information characteristics of correspondence carry out in a preset time period, and different user behavior features adopts different preset time period to add up.Described preset time period length is based on the available time of described user behavior, and available time is longer, and preset time period is longer.
This information that the different weights of described default different user behavioural characteristic show based on user behavior is to the importance of described user, and importance is higher, and weight is larger.In one embodiment, the weight of " browse, collect, place an order, strike a bargain " from low to high.
Described Preset Time decay principle is that the time is nearer, and weight is larger, and the time is far away, and weight is less.
Above-mentioned information jettison system 20 arranges different weights by different user behavior features, and successively different weighted values is arranged to the generation time of same user behavior feature, the probable value calculated more can embody the current interest-degree of user, thus the information more accurately that realizes is thrown in.
In other embodiments, described information jettison system can also comprise screening module, and for passing through information characteristics, information throws in window feature, and at least one in user's static nature is screened multiple presupposed information.Info push module 207 for selecting presupposed information from the presupposed information after screening, and is pushed to described user.
For advertisement as impression information, refer to Fig. 3, it is describe from macroscopic perspective the prediction flow process Organization Chart clicking probability, and it comprises:
Feature gathers layer
This layer data comprises following characteristics:
Characteristic of advertisement: comprise the classification of advertisement, title, unit price, and the information such as advertiser.
Advertisement position feature: comprise advertisement position specification, affiliated page classification, can the information such as advertising listing be thrown
User's static nature: comprise the relatively stable fixed informations such as age of user, sex, region, occupation.
User behavior feature: mainly refer to that the behavior of user on each platform gathers.This feature extensibility is strong, from resource, can be the independent characteristic of each platform resource, also can be multiple resource data gather feature; From the time, have by the hour, daily, the tabulate statistics feature of each granularity such as monthly.General time span is longer, and data are more stable, and the time is nearer, and data are more timely, and accuracy that is ageing and data taken into account by this layer of needs.
Prediction logic layer
It is by a kind of machine learning method supporting large-scale data amount, and feature based gathers the feature of layer, the click probability of prediction user each candidate locations on different advertisement position, exports top N advertising listing.Contain advertisement in this top N list and click probability, during input, can throw according to click probability carousel.
Advertisement putting layer
For accepting external user, ad spot information, by advertisement position resource, user's directed information, does elementary filtration, calculates candidate locations list, is transferred to prediction logic layer.And receive the top N advertising listing that prediction logic layer returns, according to regular broadcast advertisement.
Refer to Fig. 4, it gathers the integration flow process of user behavior feature in layer for the feature shown in Fig. 3, and main modular is as follows:
Platform Resource Layer
Store the source data of each platform self, each platform is can parallel expansion, and usual data source is more, the user that can represent and the match information of advertisement more, the click probability calculated is more accurate.
Platform time varying characteristic layer
From the feature that Platform Resource Layer screens, for advertisement launching platform feature, to the screening filtering of some noise data, raw data is done to the operations such as data screening, cleaning, conversion, statistical summaries.
Each platform features is analyzed, for different features, adopts different time granularity statistical summaries.Such as, unique user at least needs N days granularities just meaningful to the clicking rate of certain advertisement, and different sexes user can arrive a minute granularity to the clicking rate of different advertisement.Again such as, due to platform to women's dress to browse click volume large, as long as to the women's dress combined data of 1 day, and more day data could stably expressed user interest possibly to camera.
Multi-platform feature clustering layer
Comprehensive multiple platform resource feature, calculates the feature that level is higher.Such as gather the data browsed, click of the platforms such as multiple website, electric business, different time granularity, gather the shot and long term interest of this user to women's dress.Because each platform data amount is limited, the behavior of each user on single platform not can completely embodies user interest, and gathers multi-platform data, can more comprehensive reflection user interest.
Time decay function layer
Interval according to selected time change, in this interval, do time attenuation transform.Such as a user clicked certain advertisement before 1 minute, and its importance degree is higher than the click before 10 minutes, because the time more closely more can reflect the up-to-date interest of user, just needed to recommend more relevant advertisement.We use the multiple life cycle functions such as logistic selective, and the time is nearer, and weight is larger, and the time is far away, and weight is less, thus reach with becoming in window for the moment, the object of time decay.
User behavior characteristic layer
This layer gathers the submodule of layer as feature, become when summarizing each platform oneself many that behavioral data and multiple platform gather many time become cluster data, often kind of data all contain many latitudes feature of Different periods expansion.
The above embodiment only have expressed several embodiment of the present invention, and it describes comparatively concrete and detailed, but therefore can not be interpreted as the restriction to the scope of the claims of the present invention.It should be pointed out that for the person of ordinary skill of the art, without departing from the inventive concept of the premise, can also make some distortion and improvement, these all belong to protection scope of the present invention.Therefore, the protection domain of patent of the present invention should be as the criterion with claims.
Claims (8)
1. an information distribution method, is characterized in that, comprises the steps:
The user behavior feature of counting user and the information characteristics of correspondence;
Obtain the weight of different user behavioural characteristic;
Obtain the time of described user behavior feature, determine time weighting according to the described time, the time of described user behavior is nearer, and described time weighting is larger, and the time is far away, and described time weighting is less;
According to weight and the described time weighting of described different user behavioural characteristic, determine the weighted value of the user behavior feature of the same race of the different time of same user;
According to multiple for the information characteristics of presupposed information of input and the corresponding relation of the information characteristics of described user behavior feature, and the weighted value of user behavior feature described in each, calculate the probable value that each presupposed information comprising described information characteristics is clicked;
According to described probable value, described presupposed information is pushed to described user;
The user behavior feature of described counting user and the information characteristics of correspondence carry out in a preset time period, and different user behavior features adopts different preset time period to add up;
Described preset time period length is based on the available time of described user behavior, and available time is longer, and preset time period is longer;
This information that the weight of described user behavior feature shows based on user behavior is to the importance of described user, and importance is higher, and weight is larger.
2. information distribution method according to claim 1, it is characterized in that, described user behavior feature comprise following at least one: browse, collect, place an order, strike a bargain, APP information, the login of game, the registration of game, game download, to post containing the microblogging of certain keyword or to forward.
3. information distribution method according to claim 1, it is characterized in that, the user behavior feature of described counting user and the information characteristics of correspondence comprise: the user ID obtaining described user, obtain and add up the user behavior feature of described user at heterogeneous networks platform and the information characteristics of correspondence according to described user ID.
4. information distribution method according to claim 1, is characterized in that, describedly according to described probable value, described presupposed information is pushed to described user and comprises: at least one presupposed information before the highest for probable value is pushed to described user.
5. an information jettison system, is characterized in that, it comprises: feature acquisition module, Weight Acquisition module, time weighting acquisition module, assignment module, probability evaluation entity and info push module,
Described feature acquisition module is used for the user behavior feature of counting user and the information characteristics of correspondence;
Described Weight Acquisition module is for obtaining the weight of different user behavioural characteristic;
Described time weighting acquisition module, for obtaining the time of described user behavior feature, determines time weighting according to the described time, and the time of described user behavior is nearer, and described time weighting is larger, and the time is far away, and described time weighting is less;
Described assignment module is used for, according to the weight of described different user behavioural characteristic and described time weighting, determining the weighted value of the user behavior feature of the same race of the different time of same user;
Described probability evaluation entity is used for multiple for the information characteristics of presupposed information of input and the corresponding relation of the information characteristics of described user behavior feature according to what preset, and the weighted value of user behavior feature described in each, calculate each and comprise the clicked probable value of described information characteristics presupposed information;
Described info push module is used for, according to described probable value, default presupposed information is pushed to described user;
The described user behavior feature of feature acquisition module counting user and the information characteristics of correspondence carry out in a preset time period, and different user behavior features adopts different preset time period to add up;
Described preset time period length is based on the available time of described user behavior, and available time is longer, and preset time period is longer;
This information that the different weights of described default different user behavioural characteristic show based on user behavior is to the importance of described user, and importance is higher, and weight is larger.
6. information jettison system according to claim 5, it is characterized in that, described user behavior feature comprise browse, collect, place an order, strike a bargain, APP information, the login of game, the registration of game, game download, to post containing the microblogging of certain keyword or at least one in forwarding.
7. information jettison system according to claim 5, it is characterized in that, described feature acquisition module is used for the user ID first obtaining described user, then obtains according to described user ID and add up the user behavior feature of described user at heterogeneous networks platform and the information characteristics of correspondence.
8. information jettison system according to claim 5, it is characterized in that, describedly according to described probable value, described presupposed information is pushed to described user and comprises: according to described probable value extract probable value the highest before at least one presupposed information be pushed to described user.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210192864.9A CN103489117B (en) | 2012-06-12 | 2012-06-12 | Method and system for information releasing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210192864.9A CN103489117B (en) | 2012-06-12 | 2012-06-12 | Method and system for information releasing |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103489117A CN103489117A (en) | 2014-01-01 |
| CN103489117B true CN103489117B (en) | 2015-07-01 |
Family
ID=49829319
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210192864.9A Active CN103489117B (en) | 2012-06-12 | 2012-06-12 | Method and system for information releasing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103489117B (en) |
Families Citing this family (28)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105096152A (en) * | 2014-05-20 | 2015-11-25 | 阿里巴巴集团控股有限公司 | Commodity popularity-based operation execution method and device |
| CN104537502A (en) * | 2015-01-15 | 2015-04-22 | 北京嘀嘀无限科技发展有限公司 | Method and device for processing orders |
| CN105898420B (en) * | 2015-01-09 | 2019-04-12 | 阿里巴巴集团控股有限公司 | A kind of video recommendation method, device and electronic equipment |
| HK1201412A2 (en) * | 2015-02-16 | 2015-08-28 | 李應樵 | Characterized wireless signal mobile messaging method and system |
| EP3301628A4 (en) * | 2015-07-09 | 2018-04-18 | Huawei Technologies Co., Ltd. | Method and device for determining advertisement value |
| CN105208113A (en) * | 2015-08-31 | 2015-12-30 | 北京百度网讯科技有限公司 | Information pushing method and device |
| CN105260477A (en) * | 2015-11-06 | 2016-01-20 | 北京金山安全软件有限公司 | Information pushing method and device |
| CN106875201A (en) * | 2015-12-11 | 2017-06-20 | 北京国双科技有限公司 | The acquisition methods and device of the visible factor of advertisement |
| CN105677895B (en) * | 2016-02-03 | 2019-04-30 | 北京千米时代科技有限公司 | A kind of acquisition device and method of mobile Internet user behavior characteristics attribute |
| CN105843864A (en) * | 2016-03-17 | 2016-08-10 | 合网络技术（北京）有限公司 | Message pushing method and device |
| JP6682990B2 (en) * | 2016-04-26 | 2020-04-15 | 富士ゼロックス株式会社 | Redirection device, printing system and program |
| CN108255856A (en) * | 2016-12-29 | 2018-07-06 | 阿里巴巴集团控股有限公司 | A kind of information object dissemination method and device |
| CN106886915B (en) * | 2017-01-17 | 2020-07-28 | 华南理工大学 | Advertisement click estimation method based on time attenuation sampling |
| WO2018148894A1 (en) * | 2017-02-16 | 2018-08-23 | 深圳大学 | Time-dependent reputation evaluation algorithm based on scoring network |
| CN108665329B (en) * | 2017-03-29 | 2024-09-24 | 北京京东尚科信息技术有限公司 | Commodity recommendation method based on user browsing behavior |
| CN107656918B (en) * | 2017-05-10 | 2019-07-05 | 平安科技（深圳）有限公司 | Obtain the method and device of target user |
| CN109214842B (en) * | 2017-06-30 | 2020-07-28 | 北京金山安全软件有限公司 | Information popularization method, device and equipment |
| CN107590691B (en) * | 2017-09-06 | 2021-01-15 | 晶赞广告（上海）有限公司 | Information publishing method and device, storage medium and terminal |
| CN109840782B (en) * | 2017-11-24 | 2023-03-24 | 腾讯科技（深圳）有限公司 | Click rate prediction method, device, server and storage medium |
| CN108197288B (en) * | 2018-01-16 | 2020-09-01 | 平安好房（上海）电子商务有限公司 | House source pushing method, device, equipment and medium according to user characteristic preference |
| CN108256971A (en) * | 2018-01-16 | 2018-07-06 | 平安好房（上海）电子商务有限公司 | User's source of houses portrait update method, device, equipment and storage medium |
| CN109189812A (en) * | 2018-08-14 | 2019-01-11 | 中国平安人寿保险股份有限公司 | A kind of index querying method, computer readable storage medium and server |
| CN109801107A (en) * | 2019-01-17 | 2019-05-24 | 有米科技股份有限公司 | A kind of advertisement liveness predictor method, device, server and storage medium |
| CN112150179B (en) * | 2019-06-28 | 2024-04-09 | 京东科技控股股份有限公司 | Information pushing method and device |
| CN112214306A (en) * | 2019-07-11 | 2021-01-12 | 北京字节跳动网络技术有限公司 | Content display weight value calculation method and device, electronic equipment and computer readable storage medium |
| CN111242689B (en) * | 2020-01-14 | 2023-11-10 | 喜大（上海）网络科技有限公司 | Advertisement putting method, advertisement putting device, server and storage medium |
| CN112417275A (en) * | 2020-11-17 | 2021-02-26 | 北京三快在线科技有限公司 | Information providing method, device storage medium and electronic equipment |
| CN112989212B (en) * | 2021-05-19 | 2021-08-10 | 腾讯科技（深圳）有限公司 | Media content recommendation method, device and equipment and computer storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101960444A (en) * | 2008-02-27 | 2011-01-26 | 雅虎公司 | Event communication platform for mobile device users |
| CN102236867A (en) * | 2011-08-15 | 2011-11-09 | 悠易互通（北京）广告有限公司 | Cloud computing-based audience behavioral analysis advertisement targeting system |
| CN102332006A (en) * | 2011-08-03 | 2012-01-25 | 百度在线网络技术（北京）有限公司 | Information push control method and device |
| CN102346899A (en) * | 2011-10-08 | 2012-02-08 | 亿赞普(北京)科技有限公司 | Method and device for predicting advertisement click rate based on user behaviors |
| CN102394798A (en) * | 2011-11-16 | 2012-03-28 | 北京交通大学 | Multi-feature based prediction method of propagation behavior of microblog information and system thereof |
-
2012
- 2012-06-12 CN CN201210192864.9A patent/CN103489117B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101960444A (en) * | 2008-02-27 | 2011-01-26 | 雅虎公司 | Event communication platform for mobile device users |
| CN102332006A (en) * | 2011-08-03 | 2012-01-25 | 百度在线网络技术（北京）有限公司 | Information push control method and device |
| CN102236867A (en) * | 2011-08-15 | 2011-11-09 | 悠易互通（北京）广告有限公司 | Cloud computing-based audience behavioral analysis advertisement targeting system |
| CN102346899A (en) * | 2011-10-08 | 2012-02-08 | 亿赞普(北京)科技有限公司 | Method and device for predicting advertisement click rate based on user behaviors |
| CN102394798A (en) * | 2011-11-16 | 2012-03-28 | 北京交通大学 | Multi-feature based prediction method of propagation behavior of microblog information and system thereof |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103489117A (en) | 2014-01-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103489117B (en) | Method and system for information releasing | |
| KR102159274B1 (en) | Methods and systems for creating a data-driven attribution model for assigning attribution credit to a plurality of events | |
| CN104063801B (en) | A kind of moving advertising recommend method based on cluster | |
| CN102811371B (en) | The method, system and device that intelligent television application program is recommended | |
| KR101650117B1 (en) | Conversion path performance measures and reports | |
| CN113169886A (en) | Communication network optimization | |
| CN104065565B (en) | The method of PUSH message, server, client terminal device and system | |
| CN104834641B (en) | The processing method and related system of network media information | |
| CN106412644A (en) | Accurate advertising method and system based on smart TV player | |
| CN105045916A (en) | Mobile game recommendation system and recommendation method thereof | |
| CN101802856A (en) | Measuring a location based advertising campaign | |
| US20100146533A1 (en) | Listing Advertisement Transmitting Device and Method | |
| CN105898209A (en) | Video platform monitoring and analyzing system | |
| US20100262486A1 (en) | Network Advertisement Transmission Device and Method | |
| CN106294830A (en) | The recommendation method and device of multimedia resource | |
| CN101375306A (en) | System and method for providing user-customized event | |
| CN103686236A (en) | Method and system for recommending video resource | |
| CN102682005A (en) | Method and device for determining preference categories | |
| CN102110265A (en) | Network advertisement effect estimating method and network advertisement effect estimating system | |
| CN105160008A (en) | Method and device for locating suggested users | |
| CN106204121A (en) | Advertisement placement method based on browser and system | |
| CN103186595A (en) | Method and system for recommending audios/videos | |
| CN101464984A (en) | Advertisement publishing control system and method for publishing advertisement by using advertisement space tag | |
| CN103761228B (en) | The rank threshold of application program determines that method and rank threshold determine system | |
| CN105975537A (en) | Sorting method and device of application program |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |