CN111538846A - Third-party library recommendation method based on mixed collaborative filtering - Google Patents
Third-party library recommendation method based on mixed collaborative filtering Download PDFInfo
- Publication number
- CN111538846A CN111538846A CN202010298379.4A CN202010298379A CN111538846A CN 111538846 A CN111538846 A CN 111538846A CN 202010298379 A CN202010298379 A CN 202010298379A CN 111538846 A CN111538846 A CN 111538846A
- Authority
- CN
- China
- Prior art keywords
- application
- recommended
- party library
- party
- list
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9536—Search customisation based on social or collaborative filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
- G06F40/295—Named entity recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/088—Non-supervised learning, e.g. competitive learning
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Animal Behavior & Ethology (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a third-party library recommendation method based on mixed collaborative filtering, which comprises the following steps: obtaining feature training data sets of the application and the third-party library according to the published application and third-party library data; training by using an unsupervised learning method to obtain a theme model; extracting entities from the application and third-party database data to construct a knowledge graph and vectorizing the knowledge graph; inputting application data to be recommended into a topic model to generate an application neighbor list; obtaining a content-based scoring list of the application to be recommended by utilizing the calling information of the application to the third-party library; inputting the application to be recommended and the third-party library list to be recommended into a knowledge graph to obtain an entity vector list; calculating the similarity of the entity vectors to obtain a knowledge graph-based rating list of the applications to be recommended; and after fusion, sorting to obtain a recommendation list based on mixed recommendation. The hybrid recommendation method provided by the invention avoids the defects of a single recommendation method, effectively solves the problems of data sparseness and cold start, and improves the recommendation accuracy.
Description
Technical Field
The invention relates to a computer technology, in particular to a third-party library recommendation method based on hybrid collaborative filtering.
Background
With the rapid development of internet technology, software and applications have become an indispensable part of people's daily life. The third party library plays a crucial role in the development process of the application. They can shorten development time, improve development efficiency and improve development quality. However, with the rapid increase of third party libraries, even for experienced developers, selecting the appropriate third party library is a time consuming and laborious task. How to select a library satisfying the requirement from a large number of complex third-party libraries becomes a difficult problem in application program development. The recommendation system can effectively solve the problem of information overload and is widely applied to various scenes in a network. The current recommendation methods of the third-party library mainly include a content-based recommendation method and a collaborative filtering-based recommendation method, and the methods have the following problems:
1. the recommendation method based on the content is to find the similarity between the third-party library and the application according to the content, and then recommend the similar third-party library to the application based on the content of the application. The method can effectively solve the cold start problem and the data sparsity problem and has good interpretability, but the method cannot recommend a third party library with potential calling relation, and the feature selection, extraction and matching of the third party library are difficult points.
2. The recommendation method based on collaborative filtering is to find some applications similar to the applications to be recommended, and then recommend the third-party library called by the applications and not called by the applications to be recommended to the applications. Since this way of clustering applications requires the use of historical third party library invocation information for the applications, there are issues of data sparseness and cold start.
Disclosure of Invention
The invention aims to solve the technical problem of providing a third-party library recommendation method based on mixed collaborative filtering aiming at the defects in the prior art.
The technical scheme adopted by the invention for solving the technical problems is as follows: a third-party library recommendation method based on hybrid collaborative filtering comprises the following steps:
1) acquiring published application and third-party library data from an application and a third-party library server; the application and third party library data comprise text description information of the application and the third party library, calling information of the application to the third party library and structured semantic information of the application and the third party library;
2) preprocessing and vectorizing the text description information of the application and the third-party library in the step 1) by a natural language processing method to obtain a feature training data set of the application and the third-party library based on contents;
3) taking the content-based feature training data sets of the application and the third-party library in the step 2) as a corpus, and training by using an unsupervised learning method to obtain a theme model of the application and the third-party library;
4) constructing a calling interaction matrix of the application to the third-party library according to the calling relation of each application to each third-party library by using the calling information of the application to the third-party library in the step 1);
5) extracting entities and relations among the entities from the application obtained in the step 1) and the structured semantic information of the third party library, and storing the entities and relations into a graph database to form a knowledge graph;
6) mapping the knowledge graph obtained in the step 5) to a low-dimensional space by using a knowledge graph representation learning method to obtain vectorization representation of each entity and relationship;
7) acquiring text description information of an application to be recommended and structured information of the application to be recommended;
8) inputting the text description information of the application to be recommended, which is obtained in the step 7), into the topic model obtained in the step 3), and obtaining a content-based neighbor list of the application to be recommended through similarity comparison of the text description information;
9) according to the calling interaction information of the content-based neighbor list of the application to be recommended to the third-party library to be recommended, carrying out weighted summation and averaging according to the similarity by using a collaborative filtering method to obtain a content-based scoring list of the application to be recommended to the third-party library to be recommended;
10) inputting the structural information of the application to be recommended and the third-party library list to be recommended, which are obtained in the step 7), into a knowledge graph, and finding an entity matched with the information by an entity identification method, so as to obtain a corresponding entity vector list generated by a knowledge graph representation learning method;
11) calculating the similarity between the entity vectorization representation of the application to be recommended and the entity vectorization representation of the third-party library to be recommended, and taking the similarity as a knowledge graph-based rating list of the application to be recommended for the third-party library to be recommended;
12) fusing the content-based and collaborative-filtering scoring list of the application to be recommended, which is obtained in the step 9), for the third-party library to be recommended, and the knowledge-based scoring list of the application to be recommended, which is obtained in the step 11), for the third-party library to be recommended, so as to obtain a mixed-recommendation-based scoring list of the application to be recommended, which is based on the third-party library to be recommended;
13) and ranking the scores of the applications to be recommended to all the third-party libraries to be recommended according to a descending order, and obtaining a Top-N third-party library recommendation list of the applications to be recommended based on mixed recommendation.
According to the scheme, the pretreatment in the step 2) through a natural language processing method comprises word segmentation, stop word removal, punctuation, low-frequency vocabulary and word drying.
According to the scheme, a text vectorization method comprising TF-IDF, Doc2Bow and LDA is adopted for vectorization in the step 2).
According to the scheme, the step 4) of constructing the calling interaction matrix of the application to the third-party library specifically comprises the following steps:
classifying the calling information of the application obtained in the step 1) to the third-party library according to application IDs (identities), and generating a preference vector of each application, namely a specific biasThe good relationship is defined as: for application a and third party library l, if a calls l, then yal1, otherwise yal1, the preference vectors of the M applications to the N third-party libraries form a calling relation interaction matrix of the applications to the third-party libraries
According to the scheme, in the step 8), a content-based neighbor list of the application to be recommended is obtained, which specifically comprises the following steps:
8.1) preprocessing the text description information to be recommended and applied by a natural language method, and converting the text description information into a vector by using a text vectorization method;
8.2) inputting the obtained vector into the topic model obtained in the step 3), and comparing the similarity, taking the top k applications with the highest similarity as a neighbor list N (a) a based on the content of the application to be recommended1,a2,…,ak+, where, similarity sim (a, a)i) The following method is adopted for calculation: cosine similarity method, Euclidean distance method, Pearson correlation coefficient method.
According to the scheme, weighting summation and averaging are carried out according to the similarity in the step 9) by using a collaborative filtering method, so as to obtain a content-based score list of the application to be recommended for the third-party library to be recommended, which is specifically as follows:
according to the calling information of the application neighbor list N (a) to be recommended based on the content, and the calling information of the third-party library l to be recommended, utilizing a collaborative filtering method according to a formula:
calculating to obtain a score S of the application a to be recommended to the third-party library l to be recommended based on content and collaborative filtering1And (a, l) calculating scores of all the third party libraries in the list of the third party libraries to be recommended according to the formula to obtain a score list of the third party libraries to be recommended, wherein the score list is based on content and collaborative filtering.
According to the scheme, the calculation in the step 11) between the entity vectorization representation of the application to be recommended and the entity vectorization representation of the third-party library to be recommendedSimilarity as a knowledge graph-based score S of the application to be recommended to the third-party library to be recommended2(a, l), wherein the similarity is calculated in the following manner: cosine similarity method, Euclidean distance method, Pearson correlation coefficient method.
According to the scheme, the fusion in the step 12) adopts a mode of weighted fusion or characteristic fusion.
The invention has the following beneficial effects:
1. the invention designs a third-party library hybrid recommendation method, so that the defects of a single recommendation method are overcome, the problems of data sparseness and cold start are effectively solved, and the indexes such as recommendation accuracy and the like can be improved;
2. by utilizing a hybrid recommendation algorithm, the information of different dimensions of the application and the third-party library is fully utilized, the third-party library with a potential calling relation can be mined, and the recommendation result is prevented from being limited in a narrow range;
3. and introducing the knowledge graph, and explaining a recommendation result by using the topological structure relationship of the knowledge graph.
Drawings
The invention will be further described with reference to the accompanying drawings and examples, in which:
FIG. 1 is a flow chart of a method of an embodiment of the present invention;
FIG. 2 is a graph showing a comparison of experimental results of examples of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the following embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Some current application recommendation methods have problems such as data sparseness, cold start and the like, and can not obtain recommendations for indirectly associated third party libraries, which causes a long tail effect to a certain extent. The invention comprehensively considers the existing recommendation methods based on content and collaborative filtering, and designs a third-party library recommendation method which combines three recommendation methods, namely a knowledge graph-based method as a main method and a content and collaborative filtering-based method as an auxiliary method.
As shown in fig. 1, the third party library recommendation method based on hybrid collaborative filtering provided by the present invention includes the following steps:
step 1: the method comprises the following steps of obtaining published data of an application and a third-party library from an application server and a third-party library server, wherein the data of the application and the third-party library are divided into three parts: the unstructured text description information of the application and the third-party library is used for describing the specific functions, application scenes, use instructions and other introductions of the application and the third-party library; the method comprises the steps that calling information of a third-party library by an application is recorded in a third-party library list called by the published application; structured semantic information of the application and the third-party library, such as the type, developer, price and installation package volume of the application, and the type, label, version and other information of the third-party library;
step 2: and (3) carrying out data cleaning pretreatment on the application obtained in the step (1) and the unstructured text description information of the third-party library by using a natural language method, wherein the treatment process comprises word segmentation, removal of stop words, punctuation marks, low-frequency words, word drying and the like. Performing vectorization processing on the preprocessed data by using text vectorization methods such as TF-IDF, Doc2Bow and LDA to obtain a characteristic training data set of the application and third party library;
and step 3: training the feature training data sets of the application and the third-party library by using an unsupervised learning method to obtain theme models of the application and the third-party library for subsequent document similarity comparison;
and 4, step 4: classifying the calling information of the third-party library obtained in the step 1 according to application IDs to generate a preference vector of each application, wherein a specific preference relationship is defined as: for application a and third party library l, if a calls l, then yal1, otherwise yal1, the preference vectors of the M applications to the N third-party libraries form a calling relation interaction matrix of the applications to the third-party libraries
And 5: and (3) constructing the knowledge graph of the related field by using the application obtained in the step (1) and the self structural semantic information of the third-party library, wherein the specific construction process comprises the following steps: entity identification, entity and relation triple extraction, entity disambiguation, knowledge complementation and the like, and storing the extracted triples by using a database to obtain an available knowledge graph;
step 6: the learning method is represented by a knowledge graph, such as: TransE, TransH, TransR and the like, map the knowledge graph to a low-dimensional vector space, and represent each entity and relation in the knowledge graph as a vector;
and 7: acquiring unstructured text description information of an application to be recommended and structured information of the application from the application and a third-party library server;
and 8: preprocessing unstructured text description information of the application to be recommended by a natural language method, converting the unstructured text description information into vectors by a text vectorization method, inputting the vectors into the topic model obtained in the step 3, comparing the similarity, and taking the top k applications with the highest similarity as a neighbor list N (a) a based on the content of the application to be recommended1,a2,…,akB, here similarity sim (a, a)i) There are many measurement methods that can be selected, such as cosine similarity, Euclidean distance, Pearson correlation coefficient, etc.;
and step 9: according to the calling information of the application neighbor list N (a) to be recommended based on the content, and the calling information of the third-party library l to be recommended, utilizing a collaborative filtering method according to a formula:
calculating to obtain a score S of the application a to be recommended to the third-party library l to be recommended based on content and collaborative filtering1(a, l) calculating scores of all third party libraries in the list of the third party libraries to be recommended according to the method to obtain a score list of the applications to be recommended based on content and collaborative filtering;
step 10: inputting the structural information of the application to be recommended and the third-party library list to be recommended, which are obtained in the step (7), into a knowledge graph, and finding an entity matched with the information by an entity identification method to obtain a corresponding entity vector list generated by a knowledge graph representation learning method;
step 11: calculating the similarity between the entity vectorization representation of the application to be recommended and the entity vectorization representation of the third-party library to be recommended as the knowledge graph-based score S of the application to be recommended on the third-party library to be recommended2(a, l), also, there are various measures for similarity to choose from, and those skilled in the art can also use other existing knowledge-graph related recommendations to calculate the score;
step 12: fusing the content and collaborative filtering-based rating list of the application to be recommended, obtained in the step 9, for the third-party library to be recommended, and the knowledge graph-based rating list of the application to be recommended, obtained in the step 11, for the third-party library to be recommended, so as to obtain a hybrid recommendation-based rating list of the application to be recommended, for the third-party library to be recommended, wherein the fusion mode can be weighting fusion, feature fusion and the like;
step 13: and ranking the scores of all the third-party libraries to be recommended of the applications to be recommended in a descending order, and generating a Top-N recommendation list based on mixed recommendation.
Description of effects:
in this embodiment, in steps 2 and 8, an LDA text vectorization method is used to perform vectorization processing on the preprocessed data, in step 12, a weighted fusion method is used to fuse the content and collaborative filtering-based score list of the application to be recommended obtained in step 9 for the third-party library to be recommended and the knowledge graph-based score list of the application to be recommended obtained in step 11 for the third-party library to be recommended, and the fusion coefficient is set to 0.5.
Comparing the method of the embodiment with the existing method, the comparison result is shown in fig. 2;
setting of a comparative experiment: the experimental data set contains 5274 applications and 471 called third party library information; the comparison methods include Collaborative Filtering (CF), LDA topic model (LDA), Neural Collaborative Filtering (NCF), knowledge-graph based methods (RippleNet), and hybrid methods based on collaborative filtering and topic model (AppLibRec); the evaluation index includes: precision (Precision @ N), Recall (Recall @ N), F1 value (F1@ N), mean of average Precision (MAP @ N), and normalized loss cumulative gain (NDCG @ N).
The experimental results show that the method of the present example (TM-MKR) performed better on all the evaluation indices than all the prior comparative methods.
It will be understood that modifications and variations can be made by persons skilled in the art in light of the above teachings and all such modifications and variations are intended to be included within the scope of the invention as defined in the appended claims.
Claims (8)
1. A third-party library recommendation method based on hybrid collaborative filtering is characterized by comprising the following steps:
1) acquiring published application and third-party library data from an application and a third-party library server; the application and third party library data comprise text description information of the application and the third party library, calling information of the application to the third party library and structured semantic information of the application and the third party library;
2) preprocessing and vectorizing the text description information of the application and the third-party library in the step 1) by a natural language processing method to obtain a feature training data set of the application and the third-party library based on contents;
3) taking the content-based feature training data sets of the application and the third-party library in the step 2) as a corpus, and training by using an unsupervised learning method to obtain a theme model of the application and the third-party library;
4) constructing a calling interaction matrix of the application to the third-party library according to the calling relation of each application to each third-party library by using the calling information of the application to the third-party library in the step 1);
5) extracting entities and relations among the entities from the application obtained in the step 1) and the structured semantic information of the third party library, and storing the entities and relations into a graph database to form a knowledge graph;
6) mapping the knowledge graph obtained in the step 5) to a low-dimensional space by using a knowledge graph representation learning method to obtain vectorization representation of each entity and relationship;
7) acquiring text description information of an application to be recommended and structured information of the application to be recommended;
8) inputting the text description information of the application to be recommended, which is obtained in the step 7), into the topic model obtained in the step 3), and obtaining a content-based neighbor list of the application to be recommended through similarity comparison of the text description information;
9) according to the calling interaction information of the content-based neighbor list of the application to be recommended to the third-party library to be recommended, carrying out weighted summation according to the similarity by using a collaborative filtering method to obtain a score list of the content-based collaborative filtering of the application to be recommended to the third-party library to be recommended;
10) inputting the structural information of the application to be recommended and the third-party library list to be recommended, which are obtained in the step 7), into a knowledge graph, and finding an entity matched with the information by an entity identification method, so as to obtain a corresponding entity vector list generated by a knowledge graph representation learning method;
11) calculating the similarity between the entity vectorization representation of the application to be recommended and the entity vectorization representation of the third-party library to be recommended, and taking the similarity as a knowledge graph-based rating list of the application to be recommended for the third-party library to be recommended;
12) fusing the content-based and collaborative-filtering scoring list of the application to be recommended, which is obtained in the step 9), for the third-party library to be recommended, and the knowledge-based scoring list of the application to be recommended, which is obtained in the step 11), for the third-party library to be recommended, so as to obtain a mixed-recommendation-based scoring list of the application to be recommended, which is based on the third-party library to be recommended;
13) and ranking the scores of the applications to be recommended to all the third-party libraries to be recommended according to a descending order, and obtaining a Top-N third-party library recommendation list of the applications to be recommended based on mixed recommendation.
2. The third-party library recommendation method based on hybrid collaborative filtering according to claim 1, wherein the preprocessing in the step 2) through a natural language processing method includes word segmentation, stop word removal, punctuation and low frequency vocabulary removal, and word drying.
3. The third party library recommendation method based on hybrid collaborative filtering according to claim 1, wherein the text vectorization method including TF-IDF, Doc2Bow, LDA is adopted for the vectorization in the step 2).
4. The third-party library recommendation method based on hybrid collaborative filtering according to claim 1, wherein the step 4) is implemented by constructing a call interaction matrix of the application to the third-party library, specifically as follows:
classifying the calling information of the application obtained in the step 1) to the third-party library according to application IDs to generate a preference vector of each application, wherein a specific preference relationship is defined as: for application a and third party library l, if a calls l, then yal1, otherwise yal1, the preference vectors of the M applications to the N third-party libraries form a calling relation interaction matrix of the applications to the third-party libraries
5. The third-party library recommendation method based on hybrid collaborative filtering according to claim 1, wherein in the step 8), a content-based neighbor list of an application to be recommended is obtained, specifically as follows:
8.1) preprocessing the text description information to be recommended and applied by a natural language method, and converting the text description information into a vector by using a text vectorization method;
8.2) inputting the obtained vector into the topic model obtained in the step 3), and taking the top k applications with the highest similarity as a neighbor list N (a) ═ a of the applications to be recommended based on the content through similarity comparison1,a2,…,akIn which the similarity sim (a, a)i) The following method is adopted for calculation: cosine similarity, euclidean distance, or pearson correlation coefficient.
6. The third-party library recommendation method based on hybrid collaborative filtering according to claim 5, wherein in the step 9), a collaborative filtering method is used for weighting and summing according to similarity to obtain an average value, and a content-based score list of the application to be recommended for the third-party library to be recommended is obtained, specifically as follows:
according to the calling information of the application neighbor list N (a) to be recommended based on the content, and the calling information of the third-party library l to be recommended, utilizing a collaborative filtering method according to a formula:
calculating to obtain a score S of the application a to be recommended to the third-party library l to be recommended based on content and collaborative filtering1And (a, l) calculating scores of all the third party libraries in the list of the third party libraries to be recommended according to the formula to obtain a score list of the third party libraries to be recommended, wherein the score list is based on content and collaborative filtering.
7. The third-party library recommendation method based on hybrid collaborative filtering as claimed in claim 1, wherein the similarity between the entity-vectorized representation of the application to be recommended and the entity-vectorized representation of the third-party library to be recommended is calculated in step 11) as the knowledge-graph-based score S of the application to be recommended for the third-party library to be recommended2(a, l), wherein the similarity is calculated in the following manner: cosine similarity, euclidean distance, or pearson correlation coefficient.
8. The third-party library recommendation method based on hybrid collaborative filtering according to claim 1, wherein the fusion in step 12) is performed by means of weighted fusion or feature fusion.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010298379.4A CN111538846A (en) | 2020-04-16 | 2020-04-16 | Third-party library recommendation method based on mixed collaborative filtering |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010298379.4A CN111538846A (en) | 2020-04-16 | 2020-04-16 | Third-party library recommendation method based on mixed collaborative filtering |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111538846A true CN111538846A (en) | 2020-08-14 |
Family
ID=71978571
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010298379.4A Pending CN111538846A (en) | 2020-04-16 | 2020-04-16 | Third-party library recommendation method based on mixed collaborative filtering |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111538846A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112612973A (en) * | 2020-12-31 | 2021-04-06 | 重庆邮电大学 | Personalized intelligent clothing matching recommendation method combining knowledge graph |
| CN112749272A (en) * | 2021-01-19 | 2021-05-04 | 国家电网有限公司 | Intelligent new energy planning text recommendation method for unstructured data |
| CN112800326A (en) * | 2021-01-18 | 2021-05-14 | 吉林大学 | Improved Ripp-MKR recommendation method combining multitask learning and knowledge graph |
| CN114265982A (en) * | 2021-12-22 | 2022-04-01 | 北京中科闻歌科技股份有限公司 | Intelligent recommendation method based on knowledge graph |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104301323A (en) * | 2014-10-23 | 2015-01-21 | 中国科学院大学 | Method for third-party application balancing personalized service and user privacy information safety |
| CN104572797A (en) * | 2014-05-12 | 2015-04-29 | 深圳市智搜信息技术有限公司 | Individual service recommendation system and method based on topic model |
| US9129225B2 (en) * | 2011-05-25 | 2015-09-08 | Nokia Technologies Oy | Method and apparatus for providing rule-based recommendations |
| WO2016077255A1 (en) * | 2014-11-12 | 2016-05-19 | Microsoft Technology Licensing, Llc | Hybrid explanations in collaborative filter based recommendation system |
| US20170061286A1 (en) * | 2015-08-27 | 2017-03-02 | Skytree, Inc. | Supervised Learning Based Recommendation System |
| US20180165554A1 (en) * | 2016-12-09 | 2018-06-14 | The Research Foundation For The State University Of New York | Semisupervised autoencoder for sentiment analysis |
| CN108345702A (en) * | 2018-04-10 | 2018-07-31 | 北京百度网讯科技有限公司 | Entity recommends method and apparatus |
| CN109522474A (en) * | 2018-10-19 | 2019-03-26 | 上海交通大学 | Recommended method based on interaction sequence data mining depth user's similitude |
| CN109857835A (en) * | 2018-12-28 | 2019-06-07 | 北京红山瑞达科技有限公司 | A kind of adaptive network security knowledge assessment method based on cognitive diagnosis theory |
| CN110377845A (en) * | 2019-07-24 | 2019-10-25 | 湘潭大学 | Collaborative filtering recommending method based on the semi-supervised LDA in section |
-
2020
- 2020-04-16 CN CN202010298379.4A patent/CN111538846A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9129225B2 (en) * | 2011-05-25 | 2015-09-08 | Nokia Technologies Oy | Method and apparatus for providing rule-based recommendations |
| CN104572797A (en) * | 2014-05-12 | 2015-04-29 | 深圳市智搜信息技术有限公司 | Individual service recommendation system and method based on topic model |
| CN104301323A (en) * | 2014-10-23 | 2015-01-21 | 中国科学院大学 | Method for third-party application balancing personalized service and user privacy information safety |
| WO2016077255A1 (en) * | 2014-11-12 | 2016-05-19 | Microsoft Technology Licensing, Llc | Hybrid explanations in collaborative filter based recommendation system |
| US20170061286A1 (en) * | 2015-08-27 | 2017-03-02 | Skytree, Inc. | Supervised Learning Based Recommendation System |
| US20180165554A1 (en) * | 2016-12-09 | 2018-06-14 | The Research Foundation For The State University Of New York | Semisupervised autoencoder for sentiment analysis |
| CN108345702A (en) * | 2018-04-10 | 2018-07-31 | 北京百度网讯科技有限公司 | Entity recommends method and apparatus |
| CN109522474A (en) * | 2018-10-19 | 2019-03-26 | 上海交通大学 | Recommended method based on interaction sequence data mining depth user's similitude |
| CN109857835A (en) * | 2018-12-28 | 2019-06-07 | 北京红山瑞达科技有限公司 | A kind of adaptive network security knowledge assessment method based on cognitive diagnosis theory |
| CN110377845A (en) * | 2019-07-24 | 2019-10-25 | 湘潭大学 | Collaborative filtering recommending method based on the semi-supervised LDA in section |
Non-Patent Citations (2)
| Title |
|---|
| HONGWEI WANG ET.AL: ""Multi-Task Feature Learning for Knowledge Graph Enhanced Recommendation"", 《 HTTPS://ARXIV.ORG/ABS/1901.08907》 * |
| JIAN CHEN ET.AL,IEEE: ""Knowledge Graph Enhanced Third-Party Library Recommendation for Mobile Application Development"", 《IEEE JOURNAL ARTICLE》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112612973A (en) * | 2020-12-31 | 2021-04-06 | 重庆邮电大学 | Personalized intelligent clothing matching recommendation method combining knowledge graph |
| CN112612973B (en) * | 2020-12-31 | 2022-03-22 | 重庆邮电大学 | Personalized intelligent clothing matching recommendation method combining knowledge graph |
| CN112800326A (en) * | 2021-01-18 | 2021-05-14 | 吉林大学 | Improved Ripp-MKR recommendation method combining multitask learning and knowledge graph |
| CN112749272A (en) * | 2021-01-19 | 2021-05-04 | 国家电网有限公司 | Intelligent new energy planning text recommendation method for unstructured data |
| CN114265982A (en) * | 2021-12-22 | 2022-04-01 | 北京中科闻歌科技股份有限公司 | Intelligent recommendation method based on knowledge graph |
| CN114265982B (en) * | 2021-12-22 | 2022-08-26 | 北京中科闻歌科技股份有限公司 | Intelligent recommendation method based on knowledge graph |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110162593B (en) | Search result processing and similarity model training method and device | |
| CN109101537B (en) | Multi-turn dialogue data classification method and device based on deep learning and electronic equipment | |
| CN109635273B (en) | Text keyword extraction method, device, equipment and storage medium | |
| CN110069709B (en) | Intention recognition method, device, computer readable medium and electronic equipment | |
| US20210382937A1 (en) | Image processing method and apparatus, and storage medium | |
| RU2628436C1 (en) | Classification of texts on natural language based on semantic signs | |
| CN111538846A (en) | Third-party library recommendation method based on mixed collaborative filtering | |
| CN112800170A (en) | Question matching method and device and question reply method and device | |
| CN113204952B (en) | Multi-intention and semantic slot joint identification method based on cluster pre-analysis | |
| CN110287314B (en) | Long text reliability assessment method and system based on unsupervised clustering | |
| US20230297629A1 (en) | Apparatus and method of performance matching | |
| CN113392209A (en) | Text clustering method based on artificial intelligence, related equipment and storage medium | |
| CN112256866A (en) | Text fine-grained emotion analysis method based on deep learning | |
| Liu et al. | Open intent discovery through unsupervised semantic clustering and dependency parsing | |
| CN110750646A (en) | Attribute description extracting method for hotel comment text | |
| CN114997288A (en) | Design resource association method | |
| CN112215629B (en) | Multi-target advertisement generating system and method based on construction countermeasure sample | |
| CN113157918A (en) | Commodity name short text classification method and system based on attention mechanism | |
| CN112434533A (en) | Entity disambiguation method, apparatus, electronic device, and computer-readable storage medium | |
| CN116680363A (en) | Emotion analysis method based on multi-mode comment data | |
| CN114491079A (en) | Knowledge graph construction and query method, device, equipment and medium | |
| CN112925912B (en) | Text processing method, synonymous text recall method and apparatus | |
| CN114328800A (en) | Text processing method and device, electronic equipment and computer readable storage medium | |
| CN117828024A (en) | Plug-in retrieval method, device, storage medium and equipment | |
| CN113516094A (en) | System and method for matching document with review experts |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for 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: 20200814 |