CN103488705A - User interest model incremental update method of personalized recommendation system - Google Patents
User interest model incremental update method of personalized recommendation system Download PDFInfo
- Publication number
- CN103488705A CN103488705A CN201310403293.3A CN201310403293A CN103488705A CN 103488705 A CN103488705 A CN 103488705A CN 201310403293 A CN201310403293 A CN 201310403293A CN 103488705 A CN103488705 A CN 103488705A
- Authority
- CN
- China
- Prior art keywords
- msub
- mrow
- document
- user
- documents
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 42
- 238000004364 calculation method Methods 0.000 claims abstract description 25
- 239000013598 vector Substances 0.000 claims description 103
- 238000004422 calculation algorithm Methods 0.000 description 5
- 238000000605 extraction Methods 0.000 description 4
- 230000007246 mechanism Effects 0.000 description 3
- 230000011218 segmentation Effects 0.000 description 2
- 238000010187 selection method Methods 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004883 computer application Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012552 review Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 230000009182 swimming Effects 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/205—Parsing
- G06F40/216—Parsing using statistical methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Probability & Statistics with Applications (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a user interest model incremental update method of a personalized recommendation system. According to the basic principle of the method, the method includes storing and generating an intermediate result of calculation of a current user interest model, and performing incremental calculating on the basis of the intermediate result when the user interest model is updated. On the premise that interest information is protected from losing during updating process, the requirements that the user interest model can be updated rapidly and continuously on condition of large data amount can be met, performances of the recommendation system can be improved, and higher-quality service is provided for users.
Description
Technical Field
The invention relates to the technical field of computer application, in particular to a user interest model incremental updating method of a personalized recommendation system.
Background
The personalized recommendation system excavates objects which are potentially interesting to each user by establishing a binary relation between the user and the recommended objects and utilizing the existing selection process or similarity relation, and further carries out personalized recommendation (Liu Jian, Zhou Tao, Wan inherits macros. research progress of the personalized recommendation system [ J ]. natural science progress, 2009,19(1), 1-15.). With the diversification of user requirements, the application of the personalized recommendation system becomes wider, and the personalized recommendation system is not only used for electronic commerce, but also used for recommending webpages, documents and the like. It is necessary for the clerks and researchers to frequently review large volumes of literature. The personalized recommendation system based on the document content information learns the reading interest of the user by collecting and analyzing the interesting document content read by the user and establishes a user interest model, and a document with high matching degree is recommended to the user by comparing the matching degree of the document content and the user interest model. There are three important modules of the personalized recommendation system based on the document content information: a user interest modeling module, a recommendation object modeling module and a recommendation algorithm module, wherein the system model is shown in figure 1.
In the recommendation system based on the document content information, a user interest modeling module is a core module and is used for extracting a user interest model from an interest document read by a user and updating the interest model according to the change of the user interest. To achieve high-precision recommendations, the user interest model must be able to accurately describe the user's current interests, and updates to the interest model must be able to quickly track changes in the user's interests.
At present, two methods are mainly used for updating a user interest model, namely a time window method and a forgetting function method, wherein the time window method is used for filtering out outdated interests by using a sliding time window, and the forgetting function method is used for attenuating the weights of the interests by using a forgetting function (Feihong, wearing, Mujade, etc.. the user interest drifting method based on an optimized time window [ J ] computer engineering, 2008,34(16),210- "214"). In the literature (SHIN H., CHO S., neighboring Property Based Pattern Selection for Support vector mechanisms [ J ]. Neural Computation,2007,19(3), 816-855.), a time window method is used to update a user interest model, which uses a sliding time window to filter out outdated interests. In the literature (KEERHI S.S., SHEVADE S.K., BHATTACHARYYA, et al.A Fast Iterative New Point Algorithm for Support Vector machine classifier Design [ J ]. IEEE Transactions on Neural Networks,2000,11(1), 124-. The method comprises the following steps of (study on a Simon, updating of a user interest model and a forgetting mechanism [ J ]. microcomputer application 2011,27(7),10-11 and 69) updating the interest model according to the characteristics of an HTML document and the browsing speed of a user, and correcting the weight of a feature word by combining a forgetting factor to forget the model. Literature (li-peak, chember, swimming, ocean. adaptive user interest model based on implicit feedback [ J ] computer engineering and applications, 2008,44(9), 76-79.) separates user interests into short-term interests, which employ a time-window update mechanism, and long-term interests, which employ an update strategy based on a time-based forgetting function.
The existing user interest model updating method emphasizes how to eliminate documents deviating from the user interest from the documents in which the user is interested and add new interested documents, so that the documents used for constructing the user interest model can better reflect the current interest of the user, and the problem of the computational efficiency of updating the user interest model is ignored. With the increase of the number of the reading documents of the user, the number of the interested documents marked by the user also increases, the problem of the calculation efficiency of the user interest model updating is gradually highlighted, and the adverse effect that the model updating speed is too low to meet the requirements of the user is caused.
Disclosure of Invention
The invention aims to solve the technical problem that the prior art is insufficient, and provides a user interest model incremental updating method of a personalized recommendation system, which improves the calculation efficiency of user interest model updating on the premise of ensuring that interest information is not lost in the updating process, meets the requirement that a user interest model can be continuously and quickly updated under the condition of huge data volume, improves the performance of the personalized recommendation system, and provides higher-quality service for users.
In order to solve the technical problems, the technical scheme adopted by the invention is as follows: a user interest model incremental updating method of a personalized recommendation system comprises the following steps:
1) user interest vector space model U based on document content is constructed0；
2) Establishing the user interest vector space model U0User interest document set D0={d01,d02,...,d0mLet D = { D }1,d2,...,dnIs a set of documents to be recommended, wherein the document diThe feature vector of,(ti2,wi2),...,(tia,wia) }; wherein d is0eRepresenting the set of documents of interest to the user D0The document in (1), e =1,2, ·, m, m is the document set D of interest to the user0Total number of documents in; t is tikRepresenting the k-th term characteristic word of the document di; w is aikRepresenting a document diThe weight of the kth characteristic word; i =1,2,. n; k =1,2,. a; a represents a document diThe total number of terms of the feature words; here, the document set to be recommended is generally collected from a network or obtained from literature;
3) when the document is recommended, calculating all document feature vectors in the document set D to be recommended and the user interest vector space model U0The similarity r is recommended to be larger than the thresholdThe documents with the value alpha feed back interesting new documents to the personalized recommendation system, and the new documents are collected
The value range of the threshold value alpha is between 0 and 1, the size of the threshold value alpha is adjusted according to the needs of a user, when the user wants to obtain more recommendation results, the value of the threshold value alpha is closer to 0, and when the user wants to obtain more accurate recommendation results, the value of the threshold value alpha is closer to 1; when the outdated documents or the documents deviating from the user interest in the document set of interest of the user are selected, respectively calculating a set D0Each document feature vector and the user interest vector space model U0Selecting the documents with r' smaller than the threshold value alpha as the documents with outdated or deviated user interest, wherein the set of the documents with outdated or deviated user interest is the documents with the outdated or deviated user interest
For a document in D 'that is the new document set, f =1, 2., q, q is the total number of documents in D' that is the new document set;for the documents in the set of documents D "that are out of date or that deviate from user interest, h =1,2,.. c, c is the total number of documents in the set of documents D" that are out of date or that deviate from user interest;
4) when the user interested document set is added, the new document set D' is added to the user interested document set D0In (2), a new first set of user-interesting documents D is formed1(ii) a When the documents which are outdated or deviate from the user interest in the user interest document set are removed, the outdated or deviated from the user interest document set D '' is separated from the user interest document set D0Removing to form a new second user interested document set D2；
5) Calculating the new first order according toUser interest document set D1Central vector of
Wherein,
set D of documents of interest to said user0The feature vector of the e-th document;
the feature vector of the f document in the new document set D'; q is the total number of documents in the new document set D';
set D of documents of interest to said user0A center vector of (d); m is the document set D of interest to the user0Total number of documents in; e =1,2,. ·, m; f =1,2,. q;
calculating a new second set of user-interesting documents D according to the formula2Central vector of
Wherein,
set D of documents of interest to said user0The feature vector of the h document;
feature vectors for documents in the set of documents D ' ' that are outdated or otherwise deviating from the user's interest; c is the total number of documents in the document set D ' ' that are outdated or deviate from the user's interest;
set D of documents of interest to said user0A center vector of (d); m is the document set D of interest to the user0The total number of Chinese documents; h =1,2,. c;
6) will be provided with
Or
All dimensions are sorted from large to small according to the weight value and selected
Or
The first N dimension of the user interest vector space model U is constructed1Or U2At the same time handle
Or
Storing the information into a personalized recommendation system; wherein N is not more than
Or
The dimension of (a); using the new user interest vector space model U1Or U2Replacing U in step 1)0A new round of recommendation is made.
In the step 1), a user interest vector space model U based on document content is constructed0The method comprises the following specific steps:
1) performing feature word selection and feature word weight calculation on all documents in which the user is interested; the document feature word selection and the feature word weight can be obtained by a keyword extraction function of ICTCCLAS Chinese word segmentation software (http:// ICTCLAS. nlpir. org /), or a feature word selection method based on word frequency;
2) extracting the feature vectors of all the documents which are interested by the user to form a document feature vector set D3；
3) Computing the document feature vector set D3The center vector of (2), the document feature vector set D3The central vectors are sorted from large to small according to the weight of each dimension, and the top M dimensions are selected as a user interest vector space model U0(ii) a Wherein M does not exceed the document feature vector set D3The dimension of the central vector.
Document feature vector set D3={d31,d32,...,d3xCentral vector of }
The calculation formula of (2) is as follows:
wherein x is the document feature vector set D3The number of middle elements;
set D of feature vectors for said document3Feature vectors of the y-th document; y =1, 2.
Document D in document set D to be recommendediFeature vector and the user interest vector space model U0The calculation formula of the similarity r is as follows:
User interest document set D0The e-th document feature vector and the user interest vector space model U0The calculation formula of the similarity r' is as follows:
the basic idea of the incremental updating method of the user interest model provided by the invention is to store and generate an intermediate result in the calculation process of the current user interest model, and perform incremental calculation on the basis of the intermediate result when the user interest model is updated.
Compared with the prior art, the invention has the beneficial effects that: aiming at the problem of the efficiency of updating the user interest model of the recommendation system based on the document content information, the updating method reduces the calculation amount when the user interest model is updated on the premise of ensuring the completeness of the user information, so that the user interest model can be updated rapidly and frequently, the performance of the personalized recommendation system is improved, the user interest tracking can be rapidly realized to adapt to the change of the user interest, and higher-quality service is provided for the user.
Drawings
FIG. 1 is a system for document content information based recommendation;
FIG. 2 is a process of constructing a user interest model according to the present invention.
Detailed Description
The process of constructing a user interest vector space model based on document contents is shown in fig. 1, and firstly, feature word selection and feature word weight calculation are carried out on a document in which a user is interested to obtain a document feature vector consisting of a group of feature words and weights thereof. The document feature vector extraction method can be obtained by utilizing the feature word extraction function of ICTCCLAS Chinese word segmentation software (http:// ICTCLAS. nlpir. org /), or a feature word selection method based on word frequency. The plurality of document feature vectors constitute a document feature vector set. After the central vector of the document feature vector set is obtained through calculation, all dimensions of the central vector are sorted from large to small according to the weight, and the top N dimensions are selected as the interest model vector of the user.
The method for calculating the central vector of the document feature vector set comprises the following steps:
document collection D3={d31,d32,...,d3xDocument d2iThe feature vector of
,(t3i2,w3i2),...,(t3im,w3im) Where t is3ikRepresenting a document d3iCharacteristic word of item k, w3ikRepresenting a document d3iWeight of the k-th feature word, thatVector of center of gravity
The calculation formula is as follows:
in the formula, the document feature vectors are summed by matching feature words of each dimension, and corresponding weights are added if the feature words are the same. The first M items of the central vector after all dimensions are sorted according to the weight are the interest model U of the user, and M does not exceed the dimension of the central vector and is generally determined by the empirical value of the training sample.
Suppose the user is interested in a document of { d }1,d2,d3And the process of establishing the user interest model is shown in table 1.
TABLE 1 user interest model building Process
Central vector in table
Calculated by formula (1), here the top 5 feature terms of the center vector are selected as the user interest model U.
The specific implementation steps of the incremental updating method provided by the invention are as follows:
is provided with a U0For the user interest model which is currently established by the user, the user interest document set for establishing the user interest model is D0={d01,d02,...,d0m}. Document set D = { D = { [ D ]1,d2,...,dnIs a document to be recommended, document diThe feature vector of
,(ti2,wi2),...,(tia,wia)}。
(1) When the documents are recommended, calculating all document feature vectors and a user model U in the set D through a cosine included angle formula0Recommending the document with the similarity r larger than a threshold value alpha, feeding back an interested new document to the system after the user browses, and setting the document set as
When the outdated documents or the documents deviating from the user interest in the document set of interest of the user are selected, respectively calculating a set D0Each document feature vector and the user interest vector space model U0Selecting the documents with r' smaller than the threshold value alpha as the documents with outdated or deviated user interest, wherein the set of the documents with outdated or deviated user interest is the documents with the outdated or deviated user interest <math>
<mrow>
<msup>
<mi>D</mi>
<mrow>
<mo>′</mo>
<mo>′</mo>
</mrow>
</msup>
<mo>=</mo>
<mo>{</mo>
<msub>
<mi>d</mi>
<msub>
<mi>b</mi>
<mn>1</mn>
</msub>
</msub>
<mo>,</mo>
<msub>
<mi>d</mi>
<msub>
<mi>b</mi>
<mn>2</mn>
</msub>
</msub>
<mo>,</mo>
<mo>.</mo>
<mo>.</mo>
<mo>.</mo>
<mo>,</mo>
<msub>
<mi>d</mi>
<msub>
<mi>b</mi>
<mi>c</mi>
</msub>
</msub>
<mo>}</mo>
<mo>;</mo>
</mrow>
</math>
(2) When the user interested document set is added, the new document set D' is added to the user interested document set D0In (1), form a new user interest document set D1(ii) a When the documents which are outdated or deviate from the user interest in the user interest document set are removed, the outdated or deviated from the user interest document set D '' is separated from the user interest document set D0Removing to form a new user interested document set D2；
(3) In order to completely reserve the user interest, avoid repeated calculation and improve the algorithm performance, the system stores a calculation user interest model U in advance0Time document set D0Central vector ofTransforming the formula (1) into a formula (2) to calculate the central vector of the new interest model after adding the new document:
transforming equation (2) to equation (3) calculates the center vector of the new interest model after eliminating outdated or documents deviating from the user's interest:
(4) will be provided withAll dimensions are sorted from large to small according to weight, and the first N dimensions are selected to construct a new user interest model U1(U2) At the same time handle
) And (4) storing into the system. Using the obtained new user interest model U1(U2) Replacing U in step (1)0And carrying out new-phase recommendation.
As can be seen from equations (2) and (3), the center vector
Both appear in these two formulas. Center vector
The method is an intermediate result of the previous calculation of the user interest model, and the core of the method is to store the central vector every time the user interest model is updated
Therefore, the part of content does not need to be recalculated when the next update is carried out, and the update efficiency is improved.
Using the example in Table 2 as an example, the user interest model described in Table 2 is adding document d4When updating, set d4{ auto, 4.0}, { insurance, 3.6}, { homemade, 2.5}, { amplitude, 2.0} }, at the center vector
Is updated on the basis of the weight w of the feature word' automobile1The calculation is as shown in equation (4),
removingDocument d1When updating, for the feature word "car", its weight w2The calculation is as shown in equation (5),
and obtaining a new user interest model center vector by analogy, wherein the updating result is shown in a table 2. This example performs incremental calculations only on the basis of the number of documents of interest to the user being 3, so the improvement in computational efficiency is not significant in this example. This example is merely used to illustrate the incremental update algorithm. In practical applications, the number of documents of interest marked by the user is large, while the number of documents added or proposed is small, and the efficiency of the incremental update algorithm is more obvious.
Comparing the user interest model extraction in Table 1 with the incremental update process proposed by the present invention in Table 2, it can be found that the center vector
As an intermediate result in the process of creating or updating the user interest model last time, the incremental updating is carried out on the basis of the intermediate result, so that a large amount of vector summation work is avoided; moreover, it can be seen that the central vector obtained by the incremental updating method provided by the invention and the document directly updated by the incremental updating methodThe same extracted from the set. Generally, a document used for building a new user interest model is composed of two parts, the first part is a newly added document of interest; the second part is the part left after the original interesting documents are removed from the documents which are deviated from the current user interest, and the number of the documents in the second part is the most. The significance of the incremental updating mode provided by the invention is that the repeated calculation work of the second part of documents is avoided, so that the updating calculation amount of the user interest model is effectively reduced.
Claims (5)
1. A user interest model incremental updating method of a personalized recommendation system is characterized by comprising the following steps:
1) user interest vector space model U based on document content is constructed0；
2) Establishing the user interest vector space model U0User interest document set D0={d01,d02,...,d0mLet D = { D }1,d2,...,dnIs a set of documents to be recommended, wherein the document diThe feature vector of
,(ti2,wi2),...,(tia,wia) }; wherein d is0eRepresenting the set of documents of interest to the user D0The document in (1), e =1,2, ·, m, m is the document set D of interest to the user0Total number of documents in; t is tikRepresenting a document diThe kth term feature word; w is aikRepresenting a document diThe weight of the kth characteristic word; i =1,2,. n; k =1,2,. a; a represents a document diThe total number of terms of the feature words;
3) when the document is recommended, calculating all document feature vectors in the document set D to be recommended and the user interest vector space model U0Recommending the document with the similarity r larger than a threshold value alpha, and feeding back an interested new document to a personalized recommendation system, wherein the new document set is
(ii) a When the outdated documents or the documents deviating from the user interest in the user interest document set are selected, respectively calculating the user interest document set D0Each document feature vector and the user interest vector space model U0Selecting the documents with r' smaller than the threshold value alpha as the documents with outdated or deviated user interest, wherein the set of the documents with outdated or deviated user interest is the documents with the outdated or deviated user interest
The value range of the threshold alpha is 0-1;
for a document in D 'that is the new document set, f =1, 2., q, q is the total number of documents in D' that is the new document set;
for the documents in the set of documents D ″ that are out of date or that deviate from the user's interest, h =1,2The total number of documents in the set of documents D ' ' that are outdated or that deviate from the user's interests;
4) when the user interested document set is added, the new document set D' is added to the user interested document set D0In (2), a new first set of user-interesting documents D is formed1(ii) a Or when the documents which are outdated or deviate from the user interest in the user interest document set D are removed, the outdated or the user interest-deviating document set D '' is removed from the user interest document set D0Removing to form a new second user interested document set D2；
5) Calculating the new first set of user-interesting documents D according to1Central vector of
Wherein,
set D of documents of interest to said user0The feature vector of the e-th document;the feature vector of the f document in the new document set D';
set D of documents of interest to said user0A center vector of (d);
calculating the new second set of user-interesting documents D according to2Central vector of
Wherein,
a feature vector of the h document in the document set D ' ' that is outdated or deviating from the user's interest;
6) will be provided with
Or
All dimensions are sorted from large to small according to the weight value and selectedOr
The first N dimension of the user interest vector space model U is constructed1Or U2At the same time handle
Or
Storing the information into a personalized recommendation system; wherein N is not more than
Or
The dimension of (a); using the new user interest vector space model U1Or U2Replacement procedure
1) In (1) U0A new round of recommendation is made.
2. The method for incrementally updating the user interest model of the personalized recommendation system according to claim 1, wherein in the step 1), a user interest vector space model U based on document contents is constructed0The method comprises the following specific steps:
1) performing feature word selection and feature word weight calculation on all documents in which the user is interested;
2) extracting the feature vectors of all the documents which are interested by the user to form a document feature vector set D3；
3) Computing the document feature vectorCollection D3The center vector of (2), the document feature vector set D3The central vectors are sorted from large to small according to the weight of each dimension, and the top M dimensions are selected as a user interest vector space model U0(ii) a Wherein M does not exceed the document feature vector set D3The dimension of the central vector.
3. The method of claim 2, wherein the document feature vector set D is a set of document feature vectors3={d31,d32,...,d3xCentral vector of }
The calculation formula of (2) is as follows:
wherein x is the document feature vector set D3The number of middle elements;
set D of feature vectors for said document3The feature vector of the y-th document; y =1, 2.
4. The method for incrementally updating the user interest model of the personalized recommendation system according to any one of claims 1 to 3, wherein the document D in the document set D to be recommended isiFeature vector and the user interest vector space model U0The calculation formula of the similarity r is as follows:
wherein | | | purple hair2Representing a two-norm.
5. The method for incrementally updating the user interest model of the personalized recommendation system as recited in claim 4, wherein the set of documents D of interest of the user0The e-th document feature vector and the user interest vector space model U0The calculation formula of the similarity r' is as follows:
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310403293.3A CN103488705B (en) | 2013-09-06 | 2013-09-06 | The user interest model increment updating method of personalized recommendation system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310403293.3A CN103488705B (en) | 2013-09-06 | 2013-09-06 | The user interest model increment updating method of personalized recommendation system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103488705A true CN103488705A (en) | 2014-01-01 |
| CN103488705B CN103488705B (en) | 2016-06-22 |
Family
ID=49828931
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310403293.3A Expired - Fee Related CN103488705B (en) | 2013-09-06 | 2013-09-06 | The user interest model increment updating method of personalized recommendation system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103488705B (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104239512A (en) * | 2014-09-16 | 2014-12-24 | 电子科技大学 | Text recommendation method |
| CN104268760A (en) * | 2014-09-24 | 2015-01-07 | 同济大学 | User interest obtaining and transmitting method and system |
| CN105260481A (en) * | 2015-11-13 | 2016-01-20 | 合一网络技术(北京)有限公司 | Evaluation method and system of push list diversity |
| CN106055661A (en) * | 2016-06-02 | 2016-10-26 | 福州大学 | Multi-interest resource recommendation method based on multi-Markov-chain model |
| WO2017088587A1 (en) * | 2015-11-24 | 2017-06-01 | 华为技术有限公司 | Data processing method and device |
| CN107562912A (en) * | 2017-09-12 | 2018-01-09 | 电子科技大学 | Sina weibo event recommendation method |
| CN107635004A (en) * | 2017-09-26 | 2018-01-26 | 义乌控客科技有限公司 | A kind of personalized service method for customizing in intelligent domestic system |
| WO2018028326A1 (en) * | 2016-08-08 | 2018-02-15 | 华为技术有限公司 | Model updating method and apparatus |
| CN108446350A (en) * | 2018-03-09 | 2018-08-24 | 华中科技大学 | A kind of recommendation method based on topic model analysis and user's length interest |
| CN110287202A (en) * | 2019-05-16 | 2019-09-27 | 北京百度网讯科技有限公司 | Data-updating method, device, electronic equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101339562A (en) * | 2008-08-15 | 2009-01-07 | 北京航空航天大学 | Portal personalized recommendation service system introducing into interest model feedback and update mechanism |
| CN101477554A (en) * | 2009-01-16 | 2009-07-08 | 西安电子科技大学 | User interest based personalized meta search engine and search result processing method |
| CN102890689A (en) * | 2011-07-22 | 2013-01-23 | 北京百度网讯科技有限公司 | Method and system for building user interest model |
-
2013
- 2013-09-06 CN CN201310403293.3A patent/CN103488705B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101339562A (en) * | 2008-08-15 | 2009-01-07 | 北京航空航天大学 | Portal personalized recommendation service system introducing into interest model feedback and update mechanism |
| CN101477554A (en) * | 2009-01-16 | 2009-07-08 | 西安电子科技大学 | User interest based personalized meta search engine and search result processing method |
| CN102890689A (en) * | 2011-07-22 | 2013-01-23 | 北京百度网讯科技有限公司 | Method and system for building user interest model |
Non-Patent Citations (2)
| Title |
|---|
| INGRID ZUKERMAN AND DAVID W. ALBRECHT: "Predictive Statistical Models for User Modeling", 《USER MODELING AND USER-ADAPTED INTERACTION》, 31 December 2001 (2001-12-31), pages 5 - 18, XP008026202, DOI: doi:10.1023/A:1011175525451 * |
| 李峰，等。: "基于隐式反馈的自适应用户兴趣模型", 《计算机工程与应用》, 31 December 2008 (2008-12-31) * |
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104239512A (en) * | 2014-09-16 | 2014-12-24 | 电子科技大学 | Text recommendation method |
| CN104239512B (en) * | 2014-09-16 | 2017-06-06 | 电子科技大学 | A kind of text recommends method |
| CN104268760A (en) * | 2014-09-24 | 2015-01-07 | 同济大学 | User interest obtaining and transmitting method and system |
| CN104268760B (en) * | 2014-09-24 | 2017-06-13 | 同济大学 | A kind of user interest is obtained and transmission method and its system |
| CN105260481A (en) * | 2015-11-13 | 2016-01-20 | 合一网络技术(北京)有限公司 | Evaluation method and system of push list diversity |
| CN105260481B (en) * | 2015-11-13 | 2019-09-17 | 优酷网络技术（北京）有限公司 | A kind of multifarious evaluating method of push list and system |
| WO2017088587A1 (en) * | 2015-11-24 | 2017-06-01 | 华为技术有限公司 | Data processing method and device |
| CN106055661A (en) * | 2016-06-02 | 2016-10-26 | 福州大学 | Multi-interest resource recommendation method based on multi-Markov-chain model |
| CN106055661B (en) * | 2016-06-02 | 2017-11-17 | 福州大学 | More interest resource recommendations based on more Markov chain models |
| WO2018028326A1 (en) * | 2016-08-08 | 2018-02-15 | 华为技术有限公司 | Model updating method and apparatus |
| CN107704929A (en) * | 2016-08-08 | 2018-02-16 | 华为技术有限公司 | A kind of model update method and device |
| CN107704929B (en) * | 2016-08-08 | 2020-10-23 | 华为技术有限公司 | Model updating method and device |
| CN107562912A (en) * | 2017-09-12 | 2018-01-09 | 电子科技大学 | Sina weibo event recommendation method |
| CN107562912B (en) * | 2017-09-12 | 2021-08-27 | 电子科技大学 | Sina microblog event recommendation method |
| CN107635004A (en) * | 2017-09-26 | 2018-01-26 | 义乌控客科技有限公司 | A kind of personalized service method for customizing in intelligent domestic system |
| CN107635004B (en) * | 2017-09-26 | 2020-12-08 | 杭州控客信息技术有限公司 | Personalized service customization method in intelligent home system |
| CN108446350A (en) * | 2018-03-09 | 2018-08-24 | 华中科技大学 | A kind of recommendation method based on topic model analysis and user's length interest |
| CN108446350B (en) * | 2018-03-09 | 2020-05-19 | 华中科技大学 | Recommendation method based on topic model analysis and long and short interests of user |
| CN110287202A (en) * | 2019-05-16 | 2019-09-27 | 北京百度网讯科技有限公司 | Data-updating method, device, electronic equipment and storage medium |
| CN110287202B (en) * | 2019-05-16 | 2022-02-15 | 北京百度网讯科技有限公司 | Data updating method and device, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103488705B (en) | 2016-06-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103488705B (en) | The user interest model increment updating method of personalized recommendation system | |
| Ginelli et al. | Covariant lyapunov vectors | |
| CN109271522A (en) | Comment sensibility classification method and system based on depth mixed model transfer learning | |
| CN104573046A (en) | Comment analyzing method and system based on term vector | |
| CN108108354B (en) | Microblog user gender prediction method based on deep learning | |
| CN106096066A (en) | The Text Clustering Method embedded based on random neighbor | |
| CN109766950A (en) | A kind of industrial user's short-term load forecasting method based on form cluster and LightGBM | |
| CN102298579A (en) | Scientific and technical literature-oriented model and method for sequencing papers, authors and periodicals | |
| CN110717103B (en) | Improved collaborative filtering method based on stack noise reduction encoder | |
| CN106294325A (en) | The optimization method and device of spatial term statement | |
| CN105893362A (en) | A method for acquiring knowledge point semantic vectors and a method and a system for determining correlative knowledge points | |
| CN109800853B (en) | Matrix decomposition method and device fusing convolutional neural network and explicit feedback and electronic equipment | |
| Bi et al. | Daily runoff forecasting based on data-augmented neural network model | |
| CN109902185A (en) | A kind of water utilities field concept knowledge mapping construction method based on DBpedia | |
| Sharma et al. | A comparison of three soft computing techniques, Bayesian regression, support vector regression, and wavelet regression, for monthly rainfall forecast | |
| CN104765852A (en) | Data mining method based on fuzzy algorithm under big data background | |
| CN102521402A (en) | Text filtering system and method | |
| Ramadhan et al. | Analysis Of Warganet Comments On It Services In Mandiri Bank Using K-Nearest Neighbor (K-Nn) Algorithm Based On Itsm Criteria | |
| CN108255819A (en) | A kind of value-added tax data integration method and system based on analysis tool SPARK | |
| CN105160357A (en) | Multimodal data subspace clustering method based on global consistency and local topology | |
| Luo et al. | A comparison of three prediction models for predicting monthly precipitation in Liaoyuan city, China | |
| Alsammak et al. | An enhanced performance of K-nearest neighbor (K-NN) classifier to meet new big data necessities | |
| Jianshan et al. | Residual GM (1, 1) Model-Based Prediction Method for Chaotic Time Series. | |
| CN116738983A (en) | Word embedding method, device and equipment for performing financial field task processing by model | |
| CN113239179B (en) | Scientific research technology interest field recognition model training method, scientific and technological resource query method and device |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20160622 Termination date: 20190906 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |