CN103870978A - Intelligent push method and device based on user attribute combination difference - Google Patents
Intelligent push method and device based on user attribute combination difference Download PDFInfo
- Publication number
- CN103870978A CN103870978A CN201210570609.3A CN201210570609A CN103870978A CN 103870978 A CN103870978 A CN 103870978A CN 201210570609 A CN201210570609 A CN 201210570609A CN 103870978 A CN103870978 A CN 103870978A
- Authority
- CN
- China
- Prior art keywords
- internet user
- information
- user
- attribute information
- intelligently pushing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to an intelligent push method and device based on user attribute combination difference. The push method comprises the following steps of filling the first attribute information of an internet user according to the network behavior of the internet user so as to obtain the second attribute information of the internet user; digitizing the second attribute information according to a predetermined rule; calculating a correlation value of the internet user and a predetermined demand according to the difference and different combined strategy algorithms; judging whether to push relevant data to the internet user or not according to the correlation value. According to the intelligent pushing method, proper useful information can be pushed to a proper user; on the other hand, for an advertiser, accurate and potential users relevant to advertisements can be found, so that the advertisement putting is more efficient; besides, for a network operator, limited network resources can be more effectively applied according to the intelligent pushing method.
Description
Technical field
The present invention relates to internet arena, especially relate to a kind of otherness intelligently pushing method and device based on user property combination.
Background technology
Traditional Internet advertising business has been carried out for many years, and along with internet, especially the fast development of development of Mobile Internet technology, particularly 3G network are under the overall background of China's large scale deployment, and the advertising business based on internet will obtain development at full speed.
But traditional Internet advertising push technology cannot be to accurately identifying for identity, only can identify IP address or the online account of user's online, therefore may produce deviation for advertisement pushing object.Nowadays the mode generally adopting does not add differentiation to user exactly, take the mode advertisement of broadcast, this can allow and not need this advertisement completely or the uninterested client of this advertisement is disliked to these waste advertisements on the one hand, the effect of such advertisement is owing to not finding target customers to have a greatly reduced quality on the other hand, in addition, the advertisement pushing of such mode has brought great challenge also to the limited network bandwidth of Virtual network operator.
Therefore, Internet advertising service propelling technology has some limitations, can not be adapted to the swift and violent growth momentum of current this business completely, nor can reasonably utilize fully limited bandwidth resources by suitable advertisement pushing to may be in face of the interested target customer of this advertisement, thereby maximization network bandwidth availability ratio maximize advertising results.
Summary of the invention
In view of the understanding to above background technology, if can propose a kind of related data according to client and find the target customers of corresponding advertisement, thereby on purpose push the push technology of Internet advertising, be a kind of otherness intelligently pushing method based on user property combination, that will be highly profitable.
The understanding of the technical matters based on to above solution required for the present invention, a first aspect of the present invention has proposed a kind of otherness intelligently pushing method of user property combination, comprises the following steps:
A. according to Internet user's network behavior, the first attribute information of described Internet user is filled up, thereby obtained the second attribute information of described Internet user;
B. according to pre-defined rule by described the second attribute information digitizing;
C. calculate the correlation values of described Internet user and described predetermined demand according to otherness and various combination policing algorithm and predetermined demand; And
D. judge whether to push related data to described Internet user according to described correlation values.
In one embodiment, described the first attribute information and/or described the second attribute information comprise certainty information and uncertain information.
In one embodiment, described certainty information comprises at least one in the following: described Internet user's sex, age, occupation, educational background, birthplace and/or location.
In one embodiment, the interest that described uncertain information is described Internet user, it comprises at least one in the following: music, physical culture, tourism, internet and/or design.
In one embodiment, in described step a, will carry out filling up of described the first attribute information according to described Internet user's network behavior.
In one embodiment, described network behavior comprises viewer's sex and/or described Internet user's the location message of described Internet user's daily record.
In one embodiment, described the second attribute information digitizing solution in described step b is method of weighted mean.
In one embodiment, the formula using in the described otherness in described step c and various combination policing algorithm is:
wherein e
1+ e
2+ e
3=1, u represents the matching degree of described certainty information, and w represents the matching degree of described uncertain information, and d represents described Internet user's liveness, e
irepresent respectively certainty information, uncertain information, liveness shared proportion in described predetermined demand.
In one embodiment, described correlation values is greater than at 0.8 o'clock, and described Internet user is the accurate user under described predetermined demand.
In one embodiment, described correlation values is between 0.6 and 0.8 time, and described Internet user is the potential user under described predetermined demand.
In addition, a second aspect of the present invention has proposed a kind of intelligently pushing device of the otherness based on user property combination, comprising:
Information filling device, it is filled up the first attribute information of described Internet user according to Internet user's network behavior, thereby obtains the second attribute information of described Internet user;
The device that quantizes, its according to pre-defined rule by described the second attribute information digitizing;
Analytical equipment, it calculates the correlation values of described Internet user and described predetermined demand according to otherness and various combination policing algorithm and predetermined demand; And
Pusher, it judges whether to push related data to described Internet user according to described correlation values.
In one embodiment, described the first attribute information and/or described the second attribute information comprise certainty information and uncertain information.
In one embodiment, in described information filling device, will carry out filling up of described the first attribute information according to described Internet user's network behavior, described network behavior comprises viewer's sex and/or described Internet user's the location message of described Internet user's daily record.
In one embodiment, be method of weighted mean at described described the second attribute information digitizing solution quantizing in device.
In one embodiment, the formula using in the described otherness in described analytical equipment and various combination policing algorithm is:
wherein e
1+ e
2+ e
3=1, u represents the matching degree of described certainty information, and w represents the matching degree of described uncertain information, and d represents described Internet user's liveness, e
irepresent respectively certainty information, uncertain information, liveness shared proportion in described predetermined demand.
Can push suitable useful information for suitable user according to intelligently pushing method of the present invention and device; On the other hand, for advertiser, it can find accurate user and/or the potential user relevant to its advertisement, thereby make the input of its advertisement more efficient, in addition, for Virtual network operator, its limited Internet resources also can be by obtaining more efficiently utilization according to intelligently pushing method of the present invention.
Accompanying drawing explanation
By reading the following detailed description to non-limiting example with reference to accompanying drawing, it is more obvious that other features, objects and advantages of the present invention will become:
Fig. 1 shows the process flow diagram according to the otherness intelligently pushing method based on user property combination of the present invention; And
Fig. 2 shows the structural representation according to the otherness intelligently pushing device based on user property combination of the present invention.
Embodiment
Fig. 1 shows the process flow diagram according to the otherness intelligently pushing method based on user property combination of the present invention.
As seen from Figure 1, comprise the following steps according to intelligently pushing method of the present invention:
First, in step 110, according to Internet user's network behavior, the first attribute information of described Internet user is filled up, thereby obtained the second attribute information of described Internet user; Then, in step 120, according to pre-defined rule by described the second attribute information digitizing; Then,, in step 130, calculate the correlation values of described Internet user and described predetermined demand according to otherness and various combination policing algorithm and predetermined demand; And last, in step 140, judge whether to push related data to described Internet user according to described correlation values.
More specifically, below, lifting respectively corresponding object lesson is described above method and feature in the dependent claims.
1. filling up of userspersonal information
Userspersonal information is herein a kind of concrete form of above-mentioned the first attribute information, and one of skill in the art is appreciated that above-mentioned the first attribute information can be also other information types of Internet user.
For filling up of userspersonal information, first we know that user property comprises a lot, such as: sex, age bracket, location, occupation, schooling, income level, hobby etc., in this article, we are divided into certainty information and uncertain information user property (being userspersonal information) in this article, wherein certainty information is mainly user's personal attribute, such as sex, age, birthplace, location, occupation, educational background etc., uncertain information is mainly user's interest, such as: music, physical culture, tourism, internet, design etc.
In internet, mobile Internet, user profile is incomplete often, has some incompleteness, so just causes well evaluated and utilization of user profile, and these users are just not easy let us and find, so we need to fill up user profile.Certainty information is due to important, and most of user has, and only need to fill up small part user; For uncertain information, generally can give some labels of user, but still incomplete, so also need to fill up, so first we will have a complete uncertain information storehouse to come conveniently to label to user.
Filling up of user profile, the strategy that we take is to fill up according to user's behavior.Wherein user's behavior comprises the content that user watches or the content of delivering (daily record, video etc.), user's operating habit (click, residence time etc.) etc., and wherein user's view content plays very large effect.
Respectively to certainty and uncertainty information filling, give one example respectively below:
For filling up of certainty information, such as we do not know a user's sex, but he has seen one piece of daily record, this piece of daily record is all male user, we can tentatively conclude that this user is male user, if this piece of daily record is that reader 80% user is the male sex, can tentatively conclude that so this user is that 80% possibility is that the male sex (can introduce in detail at Part II, his sex is 0.8), for multiple user properties, we take average weighted method.
For filling up of uncertain information, the strategy that we take is the number of times that uncertain information storehouse occurs in his user behavior, according to arranging from big to small, front b the information that number of times is greater than to a is taken out, and these form user's uncertain information collection, and wherein a, b are empirical values, wherein the meaning of a be occurrence number very little, be not enough to form an interest, b is because a people's energy is limited, and uncertain information should have a threshold value.
Such as, a=10, b=6, in certain user's user behavior, uncertain information occurrence number is successively: film, tourism, football, design, internet, animation.
After filling up, user's personal information is comparatively complete, thereby is convenient to follow-up processing.For can be with the second attribute information of concrete mathematical method analysis such as userspersonal information, must be by these the second attribute information digitizings, then could facilitate follow-up algorithm process, below will introduce for example concrete how by digitized the second attribute information such as user profile.
2. the digitized processing of user profile
In step 1, user profile is divided into certainty information and uncertain information, and user profile is filled up, and user profile is carried out to digitized processing below, because the digitizing of user profile is convenient to calculate and process.
In certainty factor, such as sex, we can stipulate that women is 0, and the male sex is 1, and in step 1, a slice daily record 80% male sex's sex is just defined as 0.8 like this.
Such as the age, the deterministic age is exactly deterministic, probabilisticly still once processes, such as see 18 years old of one piece of daily record have 66,22 years old have 66, this probabilistic weighted mean value that is just defined as so, this user's age is exactly 20 years old so.
For region, if user profile is thicker, we are only sub-divided into province, and we can carry out label to province like this, and from 1 to 35, if user profile is more detailed, we can define a two-dimensional coordinate according to two-dimensional map, and we can be according to the coordinate figure definition position at place like this, value between our needed objective and user like this, can calculate with the range formula between 2 o'clock.
In uncertain factor, we give a mark to user's label, according to user's fancy grade, we carry out an arrangement, rule of thumb give a weighted value to user's hobby respectively, such as a user's interest is followed successively by internet, tourism, physical culture, these three interest indexes may be 9,8,8 (this weight index are generally decided to be 0-10) so.
The judgement of weight index:
Method has two:
First method: we are the ratio at whole uncertain collection according to some uncertain factors, determines interest index, and such as, occurrence number is greater than 30%, we remove index is that the 9...... that is decided to be of 10,20%-30% calculates down successively.
Second method:
First we we match and select a part of user (kind child user) to analyze, ratio according to some uncertain factors at whole uncertain collection, determine interest index, such as, occurrence number is greater than 30%, we remove index is 10, 20%-30% is decided to be 9, calculate successively, then according to user's behavior, be worth computing formula to be calculated with reference to google pr, that is: PR (A)=(1-d)+d (PR (t1)/C (t1)+...+PR (tn)/C (tn)), because this formula is not inventive point of the present invention, therefore do not repeat them here.
3. otherness and various combination policing algorithm
By the otherness in article step 130 and various combination policing algorithm, the second attribute information such as user profile having quantized can be converted to a correlation values particularly by means of this algorithm below, this correlation values is between 0 and 1.Particularly, in step 2, we are all information to user, comprise the processing that quantizes of certainty information and uncertain information, we according to demand, no matter be for user pushes interested demand, or throw in advertisement for advertiser below, this is all a requirements set, and we split into user's certainty information part and uncertain information part to these set.
Our otherness and various combination policing algorithm are:
Owing to may having again a lot of events in each, it is also the mode that adopts product.
That is:
After having introduced this formula, next just can the accurate user of calculative determination and potential user, we have carried out algorithm to otherness and various combination strategy and have determined, according to algorithm, we disassemble into three parts the thing of party in request, calculate respectively according to these three parts, then carry out product, calculate.
Result of calculation is the numeral of a 0-1, and we can judge and the goodness of fit of demand according to numeral, and the user that we are greater than 0.8 (empirical value) the goodness of fit is defined as accurate user, and the user of 0.6-0.8 is defined as potential user.
4. throw in the interested information of user or for accurately for advertiser throws in advertisement
Finally, require supply volume according to above-mentioned correlation values and service performance or advertiser, throw in advertisement to user, also can provide the quantity that can throw in to advertiser, be convenient to advertiser's managing advertisement expense simultaneously.
Fig. 2 shows the structural representation according to the otherness intelligently pushing device based on user property combination of the present invention.
As seen from the figure, according to the intelligently pushing device 200 of the otherness based on user property combination of the present invention, it comprises:
The device 220 that quantizes, its according to pre-defined rule by described the second attribute information digitizing;
In one embodiment, described the first attribute information and/or described the second attribute information comprise certainty information and uncertain information.Mentioned herein and certainty information and the kind of uncertain information in the description of Fig. 1, set forth particularly, do not repeat them here.
In one embodiment, in described information filling device, will carry out filling up of described the first attribute information according to described Internet user's network behavior, described network behavior comprises viewer's sex and/or described Internet user's the location message of described Internet user's daily record.Mentioned herein and certainty information and the complementing method of uncertain information in the description of Fig. 1, set forth particularly, do not repeat them here.
In one embodiment, be method of weighted mean at described described the second attribute information digitizing solution quantizing in device.The concrete example of the method was specifically set forth in the description of Fig. 1, did not repeat them here.
In one embodiment, the formula using in the described otherness in described analytical equipment and various combination policing algorithm is:
wherein e
1+ e
2+ e
3=1, u represents the matching degree of described certainty information, and w represents the matching degree of described uncertain information, and d represents described Internet user's liveness, e
irepresent respectively certainty information, uncertain information, liveness shared proportion in described predetermined demand.The specific implementation method of this formula was specifically set forth in the description of Fig. 1, did not repeat them here.
Can push suitable useful information for suitable user according to intelligently pushing method of the present invention and device; On the other hand, for advertiser, it can find accurate user and/or the potential user relevant to its advertisement, thereby make the input of its advertisement more efficient, in addition, for Virtual network operator, its limited Internet resources also can be by obtaining more efficiently utilization according to intelligently pushing method of the present invention.
Those skilled in the art will be understood that above-described embodiment is all exemplary and nonrestrictive.The different technologies feature occurring in different embodiment can combine, to obtain beneficial effect.Those skilled in the art, on the basis of research accompanying drawing, instructions and claims, will be understood that and realize the embodiment of other variations of disclosed embodiment.In claims, term " comprises " does not get rid of other devices or step; " one " does not get rid of multiple; Term " first ", " second " are for indicating title but not for representing any specific order.Any Reference numeral in claim all should not be understood to the restriction to protection domain.Some technical characterictic appears in different dependent claims and does not mean that and these technical characterictics can not be combined to obtain beneficial effect.This patent covers literal all devices and the product of going up or falling into the scope of claims in doctrine of equivalents.
Claims (15)
1. an intelligently pushing method for the otherness based on user property combination, comprises the following steps:
A. according to Internet user's network behavior, the first attribute information of described Internet user is filled up, thereby obtained the second attribute information of described Internet user;
B. according to pre-defined rule by described the second attribute information digitizing;
C. calculate the correlation values of described Internet user and described predetermined demand according to otherness and various combination policing algorithm and predetermined demand; And
D. judge whether to push related data to described Internet user according to described correlation values.
2. intelligently pushing method according to claim 1, is characterized in that, described the first attribute information and/or described the second attribute information comprise certainty information and uncertain information.
3. intelligently pushing method according to claim 2, is characterized in that, described certainty information comprises at least one in the following: described Internet user's sex, age, occupation, educational background, birthplace and/or location.
4. intelligently pushing method according to claim 2, is characterized in that, the interest that described uncertain information is described Internet user, and it comprises at least one in the following: music, physical culture, tourism, internet and/or design.
5. intelligently pushing method according to claim 1, is characterized in that, in described step a, will carry out filling up of described the first attribute information according to described Internet user's network behavior.
6. intelligently pushing method according to claim 5, is characterized in that, described network behavior comprises viewer's sex and/or described Internet user's the location message of described Internet user's daily record.
7. intelligently pushing method according to claim 1, is characterized in that, described the second attribute information digitizing solution in described step b is method of weighted mean.
8. intelligently pushing method according to claim 2, is characterized in that, the formula using in the described otherness in described step c and various combination policing algorithm is:
wherein e
1+ e
2+ e
3=1, u represents the matching degree of described certainty information, and w represents the matching degree of described uncertain information, and d represents described Internet user's liveness, e
irepresent respectively certainty information, uncertain information, liveness shared proportion in described predetermined demand.
9. intelligently pushing method according to claim 1, is characterized in that, described correlation values is greater than at 0.8 o'clock, and described Internet user is the accurate user under described predetermined demand.
10. intelligently pushing method according to claim 1, is characterized in that, described correlation values is between 0.6 and 0.8 time, and described Internet user is the potential user under described predetermined demand.
The intelligently pushing device of 11. 1 kinds of othernesses based on user property combination, comprising:
Information filling device, it is filled up the first attribute information of described Internet user according to Internet user's network behavior, thereby obtains the second attribute information of described Internet user;
The device that quantizes, its according to pre-defined rule by described the second attribute information digitizing;
Analytical equipment, it calculates the correlation values of described Internet user and described predetermined demand according to otherness and various combination policing algorithm and predetermined demand; And
Pusher, it judges whether to push related data to described Internet user according to described correlation values.
12. intelligently pushing devices according to claim 11, is characterized in that, described the first attribute information and/or described the second attribute information comprise certainty information and uncertain information.
13. intelligently pushing devices according to claim 11, it is characterized in that, in described information filling device, will carry out filling up of described the first attribute information according to described Internet user's network behavior, described network behavior comprises viewer's sex and/or described Internet user's the location message of described Internet user's daily record.
14. intelligently pushing devices according to claim 11, is characterized in that, are method of weighted mean at described described the second attribute information digitizing solution quantizing in device.
15. intelligently pushing devices according to claim 12, is characterized in that, the formula using in the described otherness in described analytical equipment and various combination policing algorithm is:
wherein e
1+ e
2+ e
3=1, u represents the matching degree of described certainty information, and w represents the matching degree of described uncertain information, and d represents described Internet user's liveness, e
irepresent respectively certainty information, uncertain information, liveness shared proportion in described predetermined demand.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210570609.3A CN103870978A (en) | 2012-12-17 | 2012-12-17 | Intelligent push method and device based on user attribute combination difference |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210570609.3A CN103870978A (en) | 2012-12-17 | 2012-12-17 | Intelligent push method and device based on user attribute combination difference |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103870978A true CN103870978A (en) | 2014-06-18 |
Family
ID=50909490
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210570609.3A Pending CN103870978A (en) | 2012-12-17 | 2012-12-17 | Intelligent push method and device based on user attribute combination difference |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103870978A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109862057A (en) * | 2017-11-30 | 2019-06-07 | 北京嘀嘀无限科技发展有限公司 | Run method for pushing, device, server and computer readable storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102103603A (en) * | 2009-12-18 | 2011-06-22 | 百度在线网络技术（北京）有限公司 | User behavior data analysis method and device |
| 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 |
| CN102799589A (en) * | 2011-05-25 | 2012-11-28 | 阿里巴巴集团控股有限公司 | Information pushing method and device |
| CN102819804A (en) * | 2011-06-07 | 2012-12-12 | 阿里巴巴集团控股有限公司 | Goods information pushing method and device |
-
2012
- 2012-12-17 CN CN201210570609.3A patent/CN103870978A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102103603A (en) * | 2009-12-18 | 2011-06-22 | 百度在线网络技术（北京）有限公司 | User behavior data analysis method and device |
| CN102799589A (en) * | 2011-05-25 | 2012-11-28 | 阿里巴巴集团控股有限公司 | Information pushing method and device |
| CN102819804A (en) * | 2011-06-07 | 2012-12-12 | 阿里巴巴集团控股有限公司 | Goods information pushing method and device |
| 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 |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109862057A (en) * | 2017-11-30 | 2019-06-07 | 北京嘀嘀无限科技发展有限公司 | Run method for pushing, device, server and computer readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR101651245B1 (en) | Conversion crediting | |
| CN109359244A (en) | A kind of recommendation method for personalized information and device | |
| US9348886B2 (en) | Formation and description of user subgroups | |
| US20160180402A1 (en) | Method for recommending products based on a user profile derived from metadata of multimedia content | |
| CN103729785B (en) | Video user gender classification method and device for method | |
| US20130086063A1 (en) | Deriving User Influences on Topics from Visual and Social Content | |
| CN106355442A (en) | Online precise advertising method and system based on big data driving | |
| Liu et al. | Using social media to infer gender composition of commuter populations | |
| CN103729360A (en) | Interest label recommendation method and system | |
| CN107526810B (en) | Method and device for establishing click rate estimation model and display method and device | |
| CN103886067A (en) | Method for recommending books through label implied topic | |
| CN106372249A (en) | Click rate estimating method and device and electronic equipment | |
| CN105608125B (en) | Information processing method and server | |
| CN107093091B (en) | Data processing method and device | |
| CN104661055A (en) | Business recommendation method and device | |
| US20130066709A1 (en) | Method of and system for determining contextually relevant advertisements to be provided to a web page | |
| US20140372202A1 (en) | Predicting performance of content items using loss functions | |
| US20110251889A1 (en) | Inventory clustering | |
| CN106127515A (en) | A kind of passenger portrait and the method and device of data analysis | |
| US9292515B1 (en) | Using follow-on search behavior to measure the effectiveness of online video ads | |
| Hong et al. | Advertising object in web videos | |
| US10699298B2 (en) | Method and system for selecting a highest value digital content | |
| Song et al. | Evaluation model of click rate of electronic commerce advertising based on fuzzy genetic algorithm | |
| CN103870978A (en) | Intelligent push method and device based on user attribute combination difference | |
| US20130246161A1 (en) | Systems and methods for optimization-aware delivery pacing adjustment in advertisement serving |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20140618 |