CN101923545B - Method for recommending personalized information - Google Patents
Method for recommending personalized information Download PDFInfo
- Publication number
- CN101923545B CN101923545B CN200910086471A CN200910086471A CN101923545B CN 101923545 B CN101923545 B CN 101923545B CN 200910086471 A CN200910086471 A CN 200910086471A CN 200910086471 A CN200910086471 A CN 200910086471A CN 101923545 B CN101923545 B CN 101923545B
- Authority
- CN
- China
- Prior art keywords
- user
- matrix
- page
- interest
- behavior
- 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.)
- Expired - Fee Related
Links
Images
Abstract
The invention discloses a method for recommending personalized information, which is suitable for the wireless Internet. The method comprises the following steps of: recording user behaviors of accessing the wireless Internet and transmitting to a server by using a behavior recording module on a mobile phone; analyzing the contents of pages accessed by a user to obtain a page set interesting to the user; analyzing the user behaviors of accessing the pages to obtain the interestingness of the user for the pages; combining the interested page set and the interestingness of the pages to establish a user interest model; and carrying out dimensional simplification to a user term matrix and calculating the similarity of term sets to generate recommended personalized information. By adopting the technical scheme, the method can automatically record and analyze the user behavior and excavate user attributes and preferences to realize the matching of the user attributes and information contents and actively recommend the information to the user.
Description
Technical field
The present invention relates to the wireless Internet technical field, relate in particular to a kind of method of recommendation of personalized information.
Background technology
Along with the fast development of Internet technology, make the information of magnanimity emerge in large numbers in face of everybody, we have got into the epoch of an information explosion already.Under this background, the user more and more is not easy therefrom to find own interested content on the one hand, also makes great deal of information nobody shows any interest on the other hand, can't be obtained by domestic consumer.Equally, the wireless Internet through the mobile phone terminal visit has realized and the intercommunication of internet, also is faced with identical problem naturally.
Search engine technique is to solve the superfluous common method of information at present; Search engine technique is included the mass data on the internet through the web crawlers traversal; Set up index and carry out the serializing storage; Utilize the front end searched page then, the keyword that the user is imported carries out Chinese and English word segmentation processing and coupling retrieval, then the qualified information content is represented according to certain sort algorithm.
There is following shortcoming in this technical scheme:
1, need initiatively inputted search keyword of user, domestic consumer can not use complicated search condition;
2, the information content that retrieval is come out under a lot of situation still is a large amount of, needs the user to carry out secondary or repeated retrieval repeatedly;
3, can't locate and the historical behavior of analysis user use preference that can not analysis user;
4, same searching key word can only be presented to the same ranking results of all users, can't corresponding service be provided to the hobby of different user.
Summary of the invention
The objective of the invention is to propose a kind of method of recommendation of personalized information, record analysis user behavior automatically, digging user attribute and preference have realized the coupling of the user property and the information content, and initiatively recommend the user.
For reaching this purpose, the present invention adopts following technical scheme:
A kind of method of recommendation of personalized information is applicable to wireless Internet, may further comprise the steps:
The behavior of the behavior record module records user capture wireless Internet on A, the mobile phone, and send to server;
B, the content of the page of user capture is analyzed, obtained the user's interest page set;
C, the behavior of user to access pages is analyzed, obtained the interest-degree of user the page;
D, said interested page set is combined with the interest-degree of the page, set up user interest model;
E, user's item matrix is carried out dimension simplify, and computational item collection similarity, the customized information of recommending produced.
Step B further may further comprise the steps:
According to user's travel log record, obtain the address of user's the browsing histories page;
From server, obtain the address corresponding page, as the data source of browsing content interest description;
From the content of pages extracting metadata, page documents is carried out text feature represent, set up the user interest matrix.
Set up collection of document D={d
1, d
2..., d
n, document d wherein
iThe employing vector space model is expressed as: d
i={ (T
1, W
1), (T
2, W
2) ..., (T
n, W
n), n is document d
iThe number of proper vector, T
iBe document d
iI proper vector, W
iBe document d
iMiddle T
iWeights.
Further comprising the steps of:
With the text participle, represent text by the characteristic speech as the dimension of vector, word frequency adopts following formula relatively:
Further comprising the steps of:
Adopt following formula to calculate the similarity of similarity between two proper vectors:
Wherein, C (X, the Y) similarity of representation page X and Y, X
iWith Y
iThe weights of expression X and Y characteristic of correspondence speech.
Among the step C, the behavior of user to access pages comprises that the user is to the browsing time of the page and the number of times of page turning/pulling scroll bar.
In the step e, adopt the singular value decomposition method that user's item matrix is carried out dimension and simplify.
In the step e, recommend collection with corresponding top-N, produce the customized information of recommending through producing the nearest-neighbors collection.
In the step e, adopt following formula calculating user's interest-degree,
Pred
U, t=u '+TS
1/2(u) S
1/2D ' (t), wherein u ' is the average assessed value of user u.
Adopted technical scheme of the present invention, through mobile phone terminal software records, analysis user behavior, tap/dip deep into user property; Sum up the user and use preference, set up the incidence relation between the user property and the information content, comprehensively use multiple proposed algorithm; Thereby can carry out personalized recommendation; Promptly recommend the different contents of information of interest separately, realized that the customized information that satisfies user preference according to user property automatically filters and commending contents, has saved user's the information retrieval time greatly to different users; Be one and solve the superfluous effective ways of present information, have important social using value.
Description of drawings
Fig. 1 is the process flow diagram of recommendation of personalized information in the specific embodiment of the invention.
Embodiment
Further specify technical scheme of the present invention below in conjunction with accompanying drawing and through embodiment.
The personalized recommendation engine is made up of three parts: collect the behavior record module of user profile, the model analysis module and the proposed algorithm module of analysis user hobby.The behavior record module is responsible for the hobby behavior of recording user; For example browse, download, subscription, question and answer, scoring, purchase etc.; The question and answer wherein and the information of scoring are relatively easily gathered; Behaviors such as yet a lot of users are reluctant that intentional system provides these information, so just need analyze user's behavior through other modes, for example browse, download, buy, subscription.The potential preference product and the fancy grade of the behavior record analysis user through these users, model analysis module that Here it is the work that will realize.The function of model analysis module can be analyzed user's behavior record, sets up the preference information that proper model is described the user.Be the proposed algorithm module at last, utilize the proposed algorithm on backstage, real-time from content library, filter out user's interest information and product is recommended.Wherein, the proposed algorithm module is the part of core the most in the commending system.
Fig. 1 is the process flow diagram of recommendation of personalized information in the specific embodiment of the invention.As shown in Figure 1, the flow process of this recommendation of personalized information may further comprise the steps:
The behavior of the behavior record module records user capture wireless Internet on step 101, the mobile phone, and send to server.
It is simple relatively that the behavior record module realizes, it is through being installed in behaviors such as the browsing of terminal software recording user on the mobile phone, subscription, and directly transfer to server and store.Through mobile phone terminal software collection user behavior inborn advantage is arranged,, can realize user's automatic login and authentication here, that is to say that we readily appreciate that whom the user who comes is.And most of users have abandoned the login of input username and password on the browser of PC, and what so obvious server was not known is whom, also just can't carry out behavior record and analysis naturally.
User interest model describe based on the web browsing content be meant the content information of user's browsing pages, it is used to content-based cluster analysis.The content information of these pages is mainly derived from the web server end; At first according to user's travel log record; Obtain sole user's browsing histories page URL, from database server, take out the corresponding web page of these URL then, as the data source that browsing content interest is described.
Compare with the structural data in the database, the web document has limited structure, even have some structures, also is to focus on form but not document content.In addition, the content of document is human employed natural language, its semanteme of computing machine intractable.These singularity in Web text message source make existing data mining technology to directly apply on it.This just need carry out pre-service to text, extracts the metadata of representing its characteristic, as the intermediate representation form of document.We adopt and to use more in recent years and effect characteristic representation preferably: vector space model (Vector Space ModelVSM) method.In VSM, text document is regarded as by one group of entry (T
1, T
2..., T
n) constitute, for each entry T
i, give certain weights W according to its significance level in article
iTherefore, the page documents of the excavation that is useful on can be used entry eigenvector { (T
1, W
1), (T
2, W
2) ..., (T
n, W
n) expression.To be a vector in the vector space with text representation, just will the text participle be represented text by these characteristic speech as the dimension of vector earlier; Initial vector representation is 0,1 form fully, that is, if occurred this speech in the text; This dimension of text vector is 1 so, otherwise is 0.These class methods can't embody the effect degree of this speech in text, so 0,1 replaced by more accurate word frequency gradually, word frequency is divided into absolute word frequency and word frequency relatively.Absolute word frequency is even the frequency of occurrences of word in text represented text; Word frequency is the word frequency of normalization relatively, and its computing method are mainly used the TF-IDF formula, have multiple TF-IDF formula at present, and we can adopt a kind of commonplace TF-1DF formula:
We the page documents that is used to excavate as a collection of document.Like this for collection of document D={d
1, d
2..., d
nIn arbitrary document d
i, adopt vector space model to be expressed as:
d
i＝{(T
1，W
1)，(T
2，W
2)，…，(T
n，W
n)}。Wherein n is document d
iThe number of proper vector, T
iBe document d
iI proper vector, W
iBe document d
iMiddle T
iWeights.
The data that adopt vector space model to represent, the necessary similarity function of selecting to calculate similarity between two eigenvectors.Method commonly used now has Euclidean distance, Manhattan distance and included angle cosine function.We adopt the included angle cosine function here.But it is different when calculating, may to run into two eigenvector length that are used for comparison, and we can adopt the method for zero-adding polishing to make both length consistent.The included angle cosine function is following:
Wherein, C (X, the Y) similarity of representation page X and Y, X
iWith Y
iThe weights of expression X and Y characteristic of correspondence speech.Page X is similar more with the Y value, and C (X, Y) value is big more; Otherwise it is then more little.
Research shows that the user much browses the interest that behavior can both reflect the user well.A lot of actions of user can both hint user's hobby, like inquiry, browsing pages and article, mark bookmark, feedback information, click the mouse, drag scroll bar, advance, retreat etc.Ask during the stop during user capture in addition, action such as access times, preservation, editor, modification also can disclose user interest.These behaviors what-the-hell reflect user's interest, and we need quantize estimation to it.
Can disclose on the surface the user to webpage P interest-degree d (P) to browse behavior a lot; But we analyze discovery, and what play a crucial role is two kinds of behaviors: the number of times v (P) (being called for short the V behavior) of the browsing time t (P) on webpage P (being called for short the T behavior) and page turning/pulling scroll bar.Reason has three: 1) behaviors such as inquiry, editor, modification must increase web page browsing time and page turning number of times, therefore can through the latter indirect obtain reflection.2) carried out the page that preservation, mark bookmark etc. move,, can repeatedly accessed later on usually and browse again, so can be presented as access times if really be that the user is concerned about.3) click the mouse the action be not considered because simple motion can not effectively disclose user interest.
In order to find T, the quantitative relationship of V and webpage interest-degree, through analyzing and experiment, we adopt the instrument of one-variable linear regression method as webpage interest modeling analysis.The linear regression analysis method is on the basis of analysis and research object variation trend, to set up function model, thus the relation of interdependence that exists between the research object.
The web browsing content analysis adopts the Web clustering method that the Web page set that the user has browsed is carried out content clustering exactly, obtains the user's interest page set; The web browsing behavioural analysis is that the behavioural information during to user's browsing pages is analyzed, and obtains the interest concentration of user to single page.The two is combined, just obtained the user's interest subject categories and reached interest-degree, promptly use the user interest model of categorize interests tree representation every type of theme.
The sparse property that the data expression method of user's item matrix of used usually collaborative filtering technology is brought has seriously restricted recommendation effect; In the bigger situation of system; It can not accurately produce recommends collection, has ignored relation potential between the data again, is necessary this matrix representation mode is done optimization; We adopt singular value decomposition, and (singular value decomposition, SVD) technology is carried out the dimension simplification to user's item matrix.
Singular value decomposition is a kind of matrix decomposition technology, and it can be 3 matrixes with the matrix decomposition of a m * n:
R＝T
0S
0D′
0，S
0＝diag(σ
1，...，σ
r)
Wherein, σ
1>=...>=σ
r>=0, T
0And D
0Be respectively the orthogonal matrix (T of m * r and n * r
0T
0'=I, D
0D
0'=I), r is order (r≤min (m, n)) of matrix R.S
0Be the diagonal matrix of a r * r, all σ
rGreater than 0 and according to the size order arrangement, be called monodrome (singularvalue).Usually for matrix R=T
0S
0D '
0, T
0, S
0And D
0It must be full rank.But singular value decomposition has an advantage, and it allows to exist the approximate matrix of a simplification.For S
0, keep k maximum monodrome, remaining is substituted with 0, like this, we just can be with S
0Be reduced to matrix that k monodrome only arranged (k＜r).Because introduced 0, can be with S
0In value be the deletion of 0 row and column, obtain a new diagonal matrix S, if matrix T
0And D
0Simplify in view of the above obtain matrix T with, the matrix R of reconstruct is arranged so
k=TSD ', R
k≈ R.All orders that singular value decomposition can generate initial matrix R equal in the matrix of k with the matrix foot recently like one.
This programme is applied to singular value decomposition in the commending system, at first is 0 sparse the mean value replacement with related column, the i.e. average assessed value of item with assessed value among the matrix R.Then the every professional etiquette model of matrix is turned to equal length, use r
IjOne r
i' the original r of replacement
Ij(r
i' be the average assessed value of the item of related column). carrying out normalized purpose is because select the user of varying number item different to the influence of similarity result of calculation; Cause a deviation easily; After standard turned to equal length, the more user of options number had reduced the influence of similarity result of calculation.Through such processing, we obtain matrix R ', and this is the input matrix of algorithm, thus, obtains our proposed algorithm:
Defeated people: matrix R ', user U, the corresponding I of set of choices with it
u
Output: correlation matrix T, S, D.
Process:
1. use singular value decomposition method split-matrix R ' to obtain matrix T
0, S
0And D
0
2. with S
0Be reduced to the matrix that dimension is k, obtain S (k＜r, r are the orders of matrix R).
3. corresponding simplification matrix T
0And D
0Obtain T, D.
4. the square root that calculates S obtains S
1/2
5. calculate two correlation matrix TS
1/2, S
1/2D '.
TS
1/2Be the matrix of m * k, what it was described is the relation of user in the k dimension space, and promptly the user is to the assessed value of one in k unit.Be appreciated that matrix, matrix S for the user
1/2D ' size is n * k, is appreciated that to be a corresponding matrix.
Adopt the vector space method to calculate similarity, analyze here to as if decompose through SVD after m * k matrix T S
1/2, we mention that its describes in the front is the relation of user in the k dimension space, because through singular value decomposition, greatly reduces its sparse property of data, can produce more accurate nearest-neighbors collection and recommend collection with corresponding top-N.
pred
u，t＝u’+TS
1/2(u)·S
1/2D’(t)
U ' is the average assessed value of user u.
The proposed algorithm of simplifying based on dimension has solved the problem of the sparse property of data preferably, simultaneously because k " n, calculation consumption has corresponding reduction, also helps solving scaling concern.The same with the collaborative filtering technology, the algorithm of simplifying based on dimension also is user oriented algorithm, and the real recommendation results that has personalized color can be provided.
This embodiment can be used the innate advantage of mobile phone terminal, automatic record analysis user behavior, and digging user attribute and preference have realized the coupling of the user property and the information content, and initiatively recommend the user.The user no longer need import keyword, the content that the general searching of also no longer need in the information of magnanimity, looking for a needle in a haystack is wanted oneself.
The above; Be merely the preferable embodiment of the present invention, but protection scope of the present invention is not limited thereto, anyly is familiar with this technological people in the technical scope that the present invention disclosed; The variation that can expect easily or replacement all should be encompassed within protection scope of the present invention.Therefore, protection scope of the present invention should be as the criterion with the protection domain of claim.
Claims (6)
1. the method for a recommendation of personalized information is applicable to wireless Internet, it is characterized in that, may further comprise the steps:
The behavior of the behavior record module records user capture wireless Internet on A, the mobile phone, and send to server;
B, the content of the page of user capture is analyzed, is obtained the user's interest page set, specifically may further comprise the steps:
According to user's travel log record, obtain the address of user's the browsing histories page,
From server, obtain the address corresponding page, as the data source of browsing content interest description,
From the content of pages extracting metadata, page documents is carried out text feature represent, set up the user interest matrix, i.e. collection of document D={d
1, d
2..., d
n, document d wherein
iThe employing vector space model is expressed as: d
i={ (T
1, W
1), (T
2, W
2) ..., (T
n, W
n), n is document d
iThe number of proper vector, T
iBe document d
iI proper vector, W
iBe document d
iMiddle T
iWeights;
C, the behavior of user to access pages is analyzed, obtained the interest-degree of user to the page, the behavior of said user to access pages is that the user is to the browsing time of the page and the number of times of page turning/pulling scroll bar;
D, said interested page set is combined with the interest-degree of the page, set up user interest model;
E, user's item matrix is carried out dimension simplify, and computational item collection similarity, the customized information of recommending produced.
2. the method for a kind of recommendation of personalized information according to claim 1 is characterized in that, step B is further comprising the steps of:
With the text participle, represent text by the characteristic speech as the dimension of vector, word frequency adopts following formula relatively:
3. the method for a kind of recommendation of personalized information according to claim 2 is characterized in that, and is further comprising the steps of:
Adopt following formula to calculate the similarity of similarity between two proper vectors:
Wherein, C (X, the Y) similarity of representation page X and Y, X
iWith Y
iThe weights of expression X and Y characteristic of correspondence speech.
4. the method for a kind of recommendation of personalized information according to claim 1 is characterized in that, in the step e, adopts the singular value decomposition method that user's item matrix is carried out dimension and simplifies.
5. based on the method for the described a kind of recommendation of personalized information of claim 4, it is characterized in that, in the step e, recommend collection with corresponding top-N, produce the customized information of recommending through producing the nearest-neighbors collection.
6. the method for a kind of recommendation of personalized information according to claim 5 is characterized in that, in the step e, adopts following formula calculating user's interest-degree,
Pred
U, t=u '+TS
1/2(u) S
1/2D ' (t), wherein u ' is the average assessed value of user u, TS
1/2Be the matrix of m * k, be used for describing the relation of user at the k dimension space, promptly the user is user's matrix to the assessed value of one in k unit, matrix S
1/2D ' size is n * k, is a corresponding matrix, and concrete calculation procedure may further comprise the steps:
Input matrix R ', user u, the corresponding I of set of choices with it
u
R ' obtains matrix T with singular value decomposition method split-matrix
0, S
0And D
0
With S
0Be reduced to the matrix that dimension is k, obtain S, k＜r wherein, r is the order of matrix R;
Corresponding simplification matrix T
0And D
0Obtain T, D;
The square root that calculates S obtains S
1/2
Calculate two correlation matrix TS
1/2, S
1/2D ';
To matrix T S
1/2The capable and matrix S of u
1/2The inner product denormalization of the t row of D ' obtains actual assessed value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910086471A CN101923545B (en) | 2009-06-15 | 2009-06-15 | Method for recommending personalized information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910086471A CN101923545B (en) | 2009-06-15 | 2009-06-15 | Method for recommending personalized information |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101923545A CN101923545A (en) | 2010-12-22 |
| CN101923545B true CN101923545B (en) | 2012-10-10 |
Family
ID=43338487
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN200910086471A Expired - Fee Related CN101923545B (en) | 2009-06-15 | 2009-06-15 | Method for recommending personalized information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101923545B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140236943A1 (en) * | 2011-11-01 | 2014-08-21 | Yahoo! Inc. | Method or system for recommending personalized content |
Families Citing this family (76)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120179717A1 (en) * | 2011-01-11 | 2012-07-12 | Sony Corporation | System and method for effectively providing entertainment recommendations to device users |
| CN102143224A (en) * | 2011-01-25 | 2011-08-03 | 张金海 | Mobile phone Internet accessing-based user behavior analysis method and device |
| CN102637170A (en) * | 2011-02-10 | 2012-08-15 | 北京百度网讯科技有限公司 | Question pushing method and system |
| CN102215300B (en) * | 2011-05-24 | 2013-11-06 | 中国联合网络通信集团有限公司 | Telecommunication service recommendation method and system |
| CN102880622A (en) * | 2011-07-15 | 2013-01-16 | 祁勇 | Method and system for determining user characteristics on internet |
| CN102937954A (en) * | 2011-08-16 | 2013-02-20 | 同程网络科技股份有限公司 | One-stop type travel information searching method |
| CN102306178A (en) * | 2011-08-25 | 2012-01-04 | 盛乐信息技术（上海）有限公司 | Video recommendation method and device |
| CN102968416A (en) * | 2011-09-01 | 2013-03-13 | 佳能株式会社 | Device and method for identifying implemented recommendation based on user intent |
| CN103034508B (en) * | 2011-10-10 | 2015-08-19 | 腾讯科技（深圳）有限公司 | Software recommendation method and system |
| CN102376063B (en) * | 2011-11-29 | 2013-09-25 | 北京航空航天大学 | Social-label-based method for optimizing personalized recommendation system |
| CN102521290A (en) * | 2011-11-29 | 2012-06-27 | 中福在线投资有限公司 | Mobile Internet lottery fan behavior metadata-based service system |
| CN103164474B (en) * | 2011-12-15 | 2016-03-30 | 中国移动通信集团贵州有限公司 | A kind of method that data service is analyzed |
| CN102402766B (en) * | 2011-12-27 | 2017-07-28 | 北京京东尚科信息技术有限公司 | A kind of user interest modeling method based on web page browsing |
| CN102591966B (en) * | 2011-12-31 | 2013-12-18 | 华中科技大学 | Filtering method of search results in mobile environment |
| CN103365842B (en) * | 2012-03-26 | 2016-08-24 | 阿里巴巴集团控股有限公司 | A kind of page browsing recommends method and device |
| CN103425670B (en) * | 2012-05-16 | 2018-11-13 | 百度在线网络技术（北京）有限公司 | A kind of method, apparatus and equipment providing a user content recommendation information |
| CN102750334B (en) * | 2012-06-01 | 2016-12-21 | 北京市农林科学院农业科技信息研究所 | A kind of method that Agricultural Information based on DM precisely pushes |
| CN103514237B (en) * | 2012-06-25 | 2018-09-04 | 深圳市易图资讯股份有限公司 | A kind of method and system obtaining user and Document personalization feature |
| CN103581278A (en) * | 2012-08-10 | 2014-02-12 | 同程网络科技股份有限公司 | User behavior analyzing method suitable for wireless client side |
| CN103678320B (en) * | 2012-09-03 | 2017-10-27 | 腾讯科技（深圳）有限公司 | The method for digging and device of the network information |
| CN103714084B (en) * | 2012-10-08 | 2018-04-03 | 腾讯科技（深圳）有限公司 | The method and apparatus of recommendation information |
| CN103729360A (en) | 2012-10-12 | 2014-04-16 | 腾讯科技（深圳）有限公司 | Interest label recommendation method and system |
| CN102982099B (en) * | 2012-11-05 | 2015-11-11 | 西安邮电大学 | A kind of personalized Parallel Word Segmentation disposal system and disposal route thereof |
| CN103514239B (en) * | 2012-11-26 | 2016-12-21 | Tcl美国研究所 | A kind of integrated user behavior and the recommendation method and system of item contents |
| CN102999634B (en) * | 2012-12-18 | 2016-07-06 | 百度在线网络技术（北京）有限公司 | Navigate recommendation method, system and cloud server based on the user of browser data |
| CN103914465A (en) * | 2012-12-31 | 2014-07-09 | 上海证大喜马拉雅网络科技有限公司 | User interest graph based intelligent customization audio listening implementation system and method |
| CN103106268B (en) * | 2013-02-01 | 2016-10-26 | 东南大学 | A kind of user preference search method |
| CN103177129B (en) * | 2013-04-19 | 2016-03-16 | 上海新数网络科技股份有限公司 | Internet real-time information recommendation prognoses system |
| CN103413243A (en) * | 2013-08-09 | 2013-11-27 | 广发银行股份有限公司 | System and method for analyzing credit card preference |
| CN103530371B (en) * | 2013-10-14 | 2017-05-24 | 李树青 | Method for realizing portable service framework |
| CN104636371B (en) * | 2013-11-11 | 2018-05-18 | 华为技术有限公司 | Information recommendation method and equipment |
| CN103617547B (en) * | 2013-12-04 | 2017-07-07 | 中国联合网络通信集团有限公司 | A kind of business recommended method and system |
| CN103761263A (en) * | 2013-12-31 | 2014-04-30 | 武汉传神信息技术有限公司 | Method for recommending information for users |
| CN104751842B (en) * | 2013-12-31 | 2019-11-15 | 科大讯飞股份有限公司 | The optimization method and system of deep neural network |
| CN104778173B (en) * | 2014-01-10 | 2020-01-10 | 腾讯科技（深圳）有限公司 | Target user determination method, device and equipment |
| CN103778329B (en) * | 2014-01-13 | 2017-01-04 | 成都国科海博信息技术股份有限公司 | A kind of construct the method that data supply value |
| CN104809127A (en) * | 2014-01-26 | 2015-07-29 | 上海联影医疗科技有限公司 | Remote education resource recommendation method and device |
| CN104850537B (en) * | 2014-02-17 | 2017-12-15 | 腾讯科技（深圳）有限公司 | The method and device screened to content of text |
| CN104951441A (en) * | 2014-03-24 | 2015-09-30 | 阿里巴巴集团控股有限公司 | Method and device for sequencing objects |
| CN103995852B (en) * | 2014-05-12 | 2018-01-09 | 百度在线网络技术（北京）有限公司 | Data processing method and device |
| CN105302807B (en) * | 2014-06-06 | 2020-01-10 | 腾讯科技（深圳）有限公司 | Method and device for acquiring information category |
| CN104063516B (en) * | 2014-07-14 | 2017-06-06 | 福州大学 | Based on the social networks rubbish filtering method that distributed matrix characteristics of decomposition is extracted |
| CN104216965B (en) * | 2014-08-21 | 2018-11-20 | 北京金山安全软件有限公司 | Information recommendation method and device |
| CN104331454A (en) * | 2014-10-30 | 2015-02-04 | 北京思特奇信息技术股份有限公司 | System and method for maintaining and retaining dormant user |
| CN105630840B (en) * | 2014-11-06 | 2019-06-28 | 阿里巴巴集团控股有限公司 | A kind of sort method and device of PUSH message |
| CN104408115B (en) * | 2014-11-25 | 2017-09-22 | 三星电子（中国）研发中心 | The heterogeneous resource based on semantic interlink recommends method and apparatus on a kind of TV platform |
| EP3026584A1 (en) | 2014-11-25 | 2016-06-01 | Samsung Electronics Co., Ltd. | Device and method for providing media resource |
| CN105095471A (en) * | 2015-08-07 | 2015-11-25 | 合肥工业大学 | Method for context sensing recommendation based on short memory |
| CN106375369B (en) * | 2016-08-18 | 2019-05-28 | 南京邮电大学 | The business recommended method of mobile Web and Collaborative Recommendation system based on user behavior analysis |
| CN106528852A (en) * | 2016-11-25 | 2017-03-22 | 盐城工学院 | Method and device for conducting redirecting by accessing data |
| CN106708802A (en) * | 2016-12-20 | 2017-05-24 | 西南石油大学 | Information recommendation method and system |
| CN106682152B (en) * | 2016-12-23 | 2020-04-14 | 西安电子科技大学 | Personalized message recommendation method |
| CN106843900B (en) * | 2017-02-09 | 2020-09-15 | 北京奇虎科技有限公司 | Information pushing method and device |
| CN107332905A (en) * | 2017-06-30 | 2017-11-07 | 广州优视网络科技有限公司 | Information-pushing method, device and server |
| CN107391694A (en) * | 2017-07-26 | 2017-11-24 | 沈丽娜 | A kind of data digging system and method |
| CN107609147B (en) * | 2017-09-20 | 2021-03-30 | 珠海金山网络游戏科技有限公司 | Method and system for automatically extracting features from log stream |
| CN107766449A (en) * | 2017-09-26 | 2018-03-06 | 杭州云赢网络科技有限公司 | Focus method for digging, device, electronic equipment and storage medium |
| CN107766484B (en) * | 2017-10-16 | 2020-09-29 | 南京师范大学 | Learning target-oriented knowledge chain recommendation method |
| CN107958070B (en) * | 2017-12-05 | 2021-11-12 | 上海电机学院 | Personalized message pushing method based on user preference |
| CN108460086B (en) * | 2018-01-22 | 2022-02-08 | 北京工业大学 | Retrieval result ordering method based on click feature reconstruction |
| CN108427723B (en) * | 2018-02-09 | 2021-08-06 | 哈尔滨工业大学深圳研究生院 | Author recommendation method and system based on clustering algorithm and local perception reconstruction model |
| CN108763367B (en) * | 2018-05-17 | 2020-07-10 | 南京大学 | Method for recommending academic papers based on deep alignment matrix decomposition model |
| CN108804619B (en) * | 2018-05-31 | 2020-11-03 | 腾讯科技（深圳）有限公司 | Interest preference prediction method, device, computer equipment and storage medium |
| CN108846062B (en) * | 2018-06-04 | 2019-08-13 | 上海市疾病预防控制中心 | Method for pushing based on users ' individualized requirement |
| CN108920537A (en) * | 2018-06-11 | 2018-11-30 | 颜彦 | The internet proposed algorithm of media service platform is melted in a kind of application |
| CN109241467A (en) * | 2018-07-30 | 2019-01-18 | 阿里巴巴集团控股有限公司 | A kind of determination method and device of page structure |
| CN109753275B (en) * | 2018-12-10 | 2022-04-12 | 东软集团股份有限公司 | Recommendation method and device for application programming interface, storage medium and electronic equipment |
| CN110688407B (en) * | 2019-09-09 | 2022-05-17 | 创新奇智(南京)科技有限公司 | Social relationship mining method |
| CN113220917B (en) * | 2020-02-06 | 2022-04-12 | 阿里巴巴集团控股有限公司 | Background map recommendation method, device and storage medium |
| CN111654518B (en) * | 2020-04-14 | 2022-05-17 | 北京奇艺世纪科技有限公司 | Content pushing method and device, electronic equipment and computer readable storage medium |
| CN111767439B (en) * | 2020-06-28 | 2023-12-15 | 百度在线网络技术（北京）有限公司 | Recommendation method, device and medium based on page classification labels |
| CN111859135B (en) * | 2020-07-22 | 2022-08-02 | 复旦大学出版社有限公司 | Anonymous user behavior mining and resource recommendation method suitable for new academic platform |
| CN112000767A (en) * | 2020-07-31 | 2020-11-27 | 深思考人工智能科技(上海)有限公司 | Text-based information extraction method and electronic equipment |
| CN112381615B (en) * | 2020-11-27 | 2022-09-02 | 华中科技大学 | Short sequence recommendation method based on user repeated behavior pattern mining |
| CN112989207B (en) * | 2021-04-27 | 2021-08-27 | 武汉卓尔数字传媒科技有限公司 | Information recommendation method and device, electronic equipment and storage medium |
| CN117670435A (en) * | 2024-02-01 | 2024-03-08 | 威海双子星软件科技有限公司 | Web application cross popularization system based on computer software and hardware integration |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1890664A (en) * | 2003-12-08 | 2007-01-03 | Iac搜索和媒体公司 | Methods and systems for conceptually organizing and presenting information |
| CN101105795A (en) * | 2006-10-27 | 2008-01-16 | 北京搜神网络技术有限责任公司 | Network behavior based personalized recommendation method and system |
| CN101256573A (en) * | 2008-04-01 | 2008-09-03 | 天津大学 | Reaction type search method and contents correlation technique based on contents relativity |
| CN101420313B (en) * | 2007-10-22 | 2011-01-12 | 北京搜狗科技发展有限公司 | Method and system for clustering customer terminal user group |
-
2009
- 2009-06-15 CN CN200910086471A patent/CN101923545B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1890664A (en) * | 2003-12-08 | 2007-01-03 | Iac搜索和媒体公司 | Methods and systems for conceptually organizing and presenting information |
| CN101105795A (en) * | 2006-10-27 | 2008-01-16 | 北京搜神网络技术有限责任公司 | Network behavior based personalized recommendation method and system |
| CN101420313B (en) * | 2007-10-22 | 2011-01-12 | 北京搜狗科技发展有限公司 | Method and system for clustering customer terminal user group |
| CN101256573A (en) * | 2008-04-01 | 2008-09-03 | 天津大学 | Reaction type search method and contents correlation technique based on contents relativity |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140236943A1 (en) * | 2011-11-01 | 2014-08-21 | Yahoo! Inc. | Method or system for recommending personalized content |
| US9607077B2 (en) * | 2011-11-01 | 2017-03-28 | Yahoo! Inc. | Method or system for recommending personalized content |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101923545A (en) | 2010-12-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101923545B (en) | Method for recommending personalized information | |
| Zhao et al. | Connecting social media to e-commerce: Cold-start product recommendation using microblogging information | |
| CN110046304B (en) | User recommendation method and device | |
| CN101641697B (en) | Related search queries for a webpage and their applications | |
| CN102929928B (en) | Multidimensional-similarity-based personalized news recommendation method | |
| CN108280114B (en) | Deep learning-based user literature reading interest analysis method | |
| Jin et al. | Web usage mining based on probabilistic latent semantic analysis | |
| CN103455487B (en) | The extracting method and device of a kind of search term | |
| CN101520784A (en) | Information issuing system and information issuing method | |
| CN107066589B (en) | Entity semantics and word frequency ordering method and device based on comprehensive knowledge | |
| CN105022827A (en) | Field subject-oriented Web news dynamic aggregation method | |
| CN104008109A (en) | User interest based Web information push service system | |
| Tan et al. | To each his own: personalized content selection based on text comprehensibility | |
| Yang | Developing an ontology-supported information integration and recommendation system for scholars | |
| Bouadjenek et al. | Persador: personalized social document representation for improving web search | |
| Duan et al. | A probabilistic mixture model for mining and analyzing product search log | |
| Bergamaschi et al. | Comparing topic models for a movie recommendation system | |
| Wang et al. | CROWN: a context-aware recommender for web news | |
| Nawazish et al. | Integrating “Random Forest” with Indexing and Query Processing for Personalized Search | |
| Peng et al. | Personalized web search using clickthrough data and web page rating | |
| Bouadjenek et al. | Personalized social query expansion using social annotations | |
| CN103309900A (en) | Personalized multidimensional document sequencing method and system | |
| Khonsha et al. | New hybrid web personalization framework | |
| Shirgave et al. | Semantically Enriched Variable Length Markov Chain Model for Analysis of User Web Navigation Sessions | |
| Chantamunee et al. | Deep autoencoder on personalized facet selection |
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 | ||
| C56 | Change in the name or address of the patentee |
Owner name: BEIJING 0080 NETWORK TECHNOLOGY CO., LTD. Free format text: FORMER NAME: BEIJING BAIFEN TONGLIAN MEDIA TECHNOLOGY CO., LTD. Owner name: OYD CAPITAL (BEIJING) CO., LTD. Free format text: FORMER NAME: BEIJING 0080 NETWORK TECHNOLOGY CO., LTD. |
|
| CP01 | Change in the name or title of a patent holder |
Address after: 3-7, room 100000, room 3 (floor five), No. 8, pioneering Road, Beijing, Haidian District, -505 Patentee after: Oriental Yuan Ding (Beijing) Cci Capital Ltd. Address before: 3-7, room 100000, room 3 (floor five), No. 8, pioneering Road, Beijing, Haidian District, -505 Patentee before: Beijing 0080 Network Technology Co.,Ltd. |
|
| CP03 | Change of name, title or address |
Address after: 3-7, room 100000, room 3 (floor five), No. 8, pioneering Road, Beijing, Haidian District, -505 Patentee after: Beijing 0080 Network Technology Co.,Ltd. Address before: 100085, No. 1, Nongda South Road, Beijing, Haidian District, Silicon Valley, No. 4, building 5 Patentee before: BEIJING LMOBILE MEDIA TECHNOLOGY Co.,Ltd. |
|
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20121010 Termination date: 20180615 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |