CN111026966B - Search recommendation ordering method based on user and product portrait and association degree of user and product portrait - Google Patents
Search recommendation ordering method based on user and product portrait and association degree of user and product portrait Download PDFInfo
- Publication number
- CN111026966B CN111026966B CN201911242968.4A CN201911242968A CN111026966B CN 111026966 B CN111026966 B CN 111026966B CN 201911242968 A CN201911242968 A CN 201911242968A CN 111026966 B CN111026966 B CN 111026966B
- Authority
- CN
- China
- Prior art keywords
- user
- similarity
- product
- commodity
- portrait
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 14
- 239000011159 matrix material Substances 0.000 claims abstract description 23
- 238000001914 filtration Methods 0.000 claims abstract description 7
- 230000006399 behavior Effects 0.000 claims description 26
- 238000004364 calculation method Methods 0.000 claims description 16
- 238000000354 decomposition reaction Methods 0.000 claims description 5
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000012935 Averaging Methods 0.000 description 1
- 230000003542 behavioural effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—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
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0623—Item investigation
- G06Q30/0625—Directed, with specific intent or strategy
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Abstract
The invention relates to a search recommendation ordering method based on user and product portraits and the association degree of the user and the product portraits, which comprises the following steps: generating features related to the product content in the user representation using a conventional recommendation system such as item-based recommendation model; using a common user behavior matrix concept in a collaborative filtering recommendation system, regarding all users who purchase a certain commodity or consult single content as an implicit user of the content, and generating a behavior matrix of the implicit user in a product portrait; generating new features to describe the similarity between the search user and the search target; and describing the similarity between the search user and the search target at multiple angles, and judging the importance degree of the features of each training data set by using an LTR algorithm. The invention fully considers the ordering mode of the user personality, so that the ordering is more targeted, thereby improving the ordering effect and the user satisfaction.
Description
Technical Field
The patent application belongs to the technical field of in-site search systems, and particularly relates to a search recommendation ordering method based on users, product portraits and association degrees of the users and the product portraits.
Background
Most of the existing in-site search engines rank search sentences and commodity descriptions based on semantic similarity by using models such as VSM, BM25 and the like when performing search ranking, or rank search results according to commodity heat and content heat. This ordering does not take into account the searcher's personality. Even if the system considers the individuality of the searcher, the characteristic generation does not relate the user image of the searcher with the product/content image of the search result, or considers the similarity degree of the two images, so the prior in-site search engine is insufficient in the aspects of improving the accuracy of sorting, providing individuality and pertinence service for the user.
Disclosure of Invention
The technical problem to be solved by the invention is to provide a search recommendation ordering method based on the user, the product portrait and the association degree of the user and the product portrait, which can improve the ordering accuracy and provide personalized and targeted service for the user.
In order to solve the problems, the invention adopts the following technical scheme:
a search recommendation ordering method based on user and product portraits and the association degree of the user and the product portraits comprises the following steps:
s1, generating comprehensive commodity characteristics of users related to product content in an original user portrait according to the attribute of all commodities purchased or referred by a certain user;
s2, generating hidden user characteristics of commodities related to user contents in the original product portrait according to the behavior characteristics of all users who have purchased or referred to a certain commodity;
s3, calculating multiple kinds of similarity between the comprehensive commodity characteristics of the user in the S1 and the attribute characteristics of corresponding commodities in the sample, and generating multiple commodity similarity results; meanwhile, carrying out multiple similarity calculation on the hidden user characteristics of the commodity in the S2 and the behavior characteristics of corresponding users in the sample to obtain multiple user similarity results;
s4, judging a plurality of commodity similarity results and a plurality of user similarity results in the S3, and calculating the importance degree of each similarity result through a learning sequencing model;
and S5, outputting the similarity result with the highest importance degree, namely outputting the commodity similarity result with the highest importance degree and the user similarity result, so as to predict and recommend purchasing behavior.
The technical scheme of the invention is further improved as follows: in S1, the comprehensive commodity characteristics of the user are generated in the original user portrait by using a traditional recommendation system.
The technical scheme of the invention is further improved as follows: the traditional recommendation system is an item-based recommendation model or a user-based recommendation model.
The technical scheme of the invention is further improved as follows: and S2, generating hidden user characteristics of the commodity in the original product portrait by using a user behavior matrix algorithm in the collaborative filtering recommendation system.
The technical scheme of the invention is further improved as follows: in S3, the similarity calculation includes vector similarity, cross phase, singular value decomposition SVD, where:
s31, adopting a cosine similarity calculation formula cosine similarity for vector similarity:
a is comprehensive commodity characteristics of a user or hidden user characteristics of the commodity, and B is attribute characteristics of corresponding commodity in a sample or behavior characteristics of corresponding user in the sample; wherein the bold meaning of A, B is different from the bold meaning of the non-bold meaning of the element in the vector, the non-bold meaning of the element in the vector, and the specific element is combined with the subscript i;
s32, intersecting phase: generating a cross item for the comprehensive commodity characteristics of the user and the attribute characteristics of the corresponding commodity in the sample, and generating a cross item for the hidden user characteristics of the commodity and the behavior characteristics of the corresponding user in the sample;
s33, SVD is carried out on the behavior characteristics of all users and the hidden user characteristics of all commodities, and the SVD is recorded as M=U ΣV T The matrix M is a matrix formed by the behavior characteristics of all users or is formed by the hidden user characteristics of all commoditiesThe matrix U obtained after SVD allows similar users, commodities with similar implicit users, to be mapped to more recent locations in the final feature space.
The technical scheme of the invention is further improved as follows: in S4, the learning ordering model is an LTR algorithm.
Due to the adoption of the technical scheme, the beneficial effects obtained by the invention are as follows:
1. combining a traditional recommendation algorithm, and fusing an evaluation index based on the traditional algorithm into the characteristics of the LTR (Learning to rank) model;
2. describing the similarity between a user and a product from different angles by using a plurality of similarity calculation modes, and learning the importance of various similarities by using a model;
3. the dimension increasing mode enables the similarity between the user and the content to be displayed in the characteristics, and enables the learning effect of the model to be enhanced in basicity;
4. the method can effectively solve the problems of cold start of new products and new contents when only the comprehensive commodity characteristics of the user are used;
5. the sorting mode fully considering the individuality of the user can lead the sorting to be more targeted, and improves the sorting effect and the user satisfaction.
Drawings
FIG. 1 is a schematic process diagram of the present invention;
FIG. 2 is a schematic diagram of class A features according to an embodiment of the present invention;
fig. 3 is a schematic diagram of a class B feature principle according to an embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to examples.
As shown in fig. 1 to 3, the invention discloses a search recommendation ordering method based on user and product portraits and the association degree of the user and the product portraits, which comprises the following steps:
s1, according to the attribute of all commodities purchased or referred by a certain user, generating comprehensive commodity characteristics of the user related to the product content in an original user portrait by using a traditional recommendation system and performing summarization calculation; the summary calculation in S1 includes one or more of weighted average, logical and, logical or. The weighted average, logical AND, logical OR, or operations performed on all the commodities once purchased by the user, because the commodity characteristics include different data types such as numerical type, discrete type, etc., there are different summary calculation modes, and the numerical type is the weighted average, and the discrete type is the logical AND, the logical OR, etc.
S2, according to the behavior characteristics of all users who have purchased or referred to a certain commodity, generating hidden user characteristics of the commodity related to the user content in the original product portrait after summarized calculation by using the user behavior matrix concept in the collaborative filtering recommendation system; the summary calculation in S1 is collaborative filtering and weighted averaging.
S3, calculating multiple kinds of similarity between the comprehensive commodity characteristics of the user in the S1 and the attribute characteristics of corresponding commodities in the sample, and generating multiple commodity similarity results; meanwhile, carrying out multiple similarity calculation on the hidden user characteristics of the commodity in the S2 and the behavior characteristics of corresponding users in the sample to obtain multiple user similarity results;
s4, judging a plurality of commodity similarity results and a plurality of user similarity results in the S3, and calculating the importance degree of each similarity result through a learning sequencing model; the Learning ordering model is the LTR algorithm (Learning to Rank).
S5, outputting the similarity result with the highest importance degree, namely outputting the commodity similarity result with the highest importance degree and the user similarity result, so as to predict and recommend purchasing behavior;
in S1, a conventional recommendation system is used to generate comprehensive commodity features of a user in an original user portrait, such as an item-based recommendation model or a user-based recommendation model.
And S2, generating hidden user characteristics of the commodity in the original product portrait by using a user behavior matrix algorithm in the collaborative filtering recommendation system.
In S3, the similarity calculation includes vector similarity, cross phase, singular value decomposition SVD, where:
s31, adopting a cosine similarity calculation formula cosine similarity for vector similarity:
a is comprehensive commodity characteristics of a user or hidden user characteristics of the commodity, and B is attribute characteristics of corresponding commodity in a sample or behavior characteristics of corresponding user in the sample; wherein the bold meaning of A, B is different from the bold meaning of the non-bold meaning of the element in the vector, the non-bold meaning of the element in the vector, and the specific element is combined with the subscript i;
s32, intersecting phase: generating a cross item for the comprehensive commodity characteristics of the user and the attribute characteristics of the corresponding commodity in the sample, and generating a cross item for the hidden user characteristics of the commodity and the behavior characteristics of the corresponding user in the sample;
s33, SVD is carried out on the behavior characteristics of all users and the hidden user characteristics of all commodities, and the SVD is recorded as M=U ΣV T The matrix M is a matrix formed by the behavior characteristics of all users or a matrix formed by the hidden user characteristics of all commodities, and the matrix U obtained after SVD is performed enables similar users and commodities with similar hidden users to be mapped to more recent positions in a final characteristic space.
SVD also decomposes matrices, but unlike feature decomposition, SVD does not require that the matrix to be decomposed be a square matrix. Assuming that our matrix a is an mx n m ×n matrix, we define the SVD of matrix a as: a=uΣv T
Where U is an m matrix, Σ is an m n matrix, all 0 except for the elements on the main diagonal, each element on the main diagonal being called singular value, and V is an n matrix. U and V are unitary matrices, i.e. satisfy U T U＝I，V T V＝I。
The following examples are set forth in detail:
1) Generating a part of characteristics related to the product content in the user portrait by using a traditional recommendation system such as item-based recommendation model, wherein the part of characteristics are hereinafter referred to as class A characteristics, and the part of characteristics can be directly calculated to be similar to the characteristics related to the product content in the product portrait, such as attributes that the user U1 has purchased commodities P1, P2, P3 and the like, so as to form class A characteristics of the user U1;
2) Using a common user behavior matrix concept in a collaborative filtering recommendation system, regarding all users who purchase a certain commodity or consult single content as an 'implicit user' of the content, generating a behavior matrix of the 'implicit user' in a product image, wherein part of characteristics are hereinafter referred to as B-class characteristics, and the similarity can be directly calculated with behavior information in the user image; such as the behavioral characteristics of user U1, user U2, user U3, etc., who have purchased P1, form the class B characteristics of the merchandise P1.
3) The following methods are respectively used for the A type and the B type features, and new features are generated to describe the similarity between the A type and the B type features:
(1) Calculating the similarity of the two types of features by using a vector similarity calculation formula, such as cosine similarity (cosine similarity);
(2) Respectively generating cross phases for the two types of features;
(3) SVD (singular value decomposition) is performed on the user behavior features and commodity class B features.
These similarity calculation methods may be used together and the result is passed as a feature into the LTR model. The term "similarity" is used herein to describe the degree of similarity between the user and the group of users who have previously purchased the item. The higher the theoretical degree of similarity, the more likely the user will purchase the item.
4) And describing the similarity between the searching user and the searching target at multiple angles, and judging the importance degree of the input features of each training data set by using an LTR algorithm (Learning to Rank), so that more effective searching recommendation is carried out, and the ordering effect and the user satisfaction degree are improved.
Claims (4)
1. A search recommendation ordering method based on user and product portraits and the association degree of the user and the product portraits is characterized by comprising the following steps:
s1, generating comprehensive commodity characteristics of users related to product content in an original user portrait according to the attribute of all commodities purchased or referred by a certain user;
s2, generating hidden user characteristics of commodities related to user contents in the original product portrait according to the behavior characteristics of all users who have purchased or referred to a certain commodity;
s3, calculating multiple kinds of similarity between the comprehensive commodity characteristics of the user in the S1 and the attribute characteristics of corresponding commodities in the sample, and generating multiple commodity similarity results; meanwhile, carrying out multiple similarity calculation on the hidden user characteristics of the commodity in the S2 and the behavior characteristics of corresponding users in the sample to obtain multiple user similarity results;
wherein, the step S3 includes:
acquiring attributes of commodities P1, P2 and P3 purchased by a user U1, and forming A-type characteristics of the user U1;
acquiring behavior characteristics of a user U1, a user U2 and a user U3 who purchase the commodity P1 to form B-type characteristics of the commodity P1;
respectively calculating the similarity of the class A features and the class B features by using a vector similarity calculation formula;
respectively generating cross phases for the A type features and the B type features;
SVD singular value decomposition is carried out on the A type features and the B type features;
s4, judging a plurality of commodity similarity results and a plurality of user similarity results in the S3, and calculating the importance degree of each similarity result through a learning sequencing model; wherein, the learning ordering model is an LTR algorithm;
s5, outputting the similarity result with the highest importance degree, so that the purchasing behavior is predicted and recommended.
2. The search recommendation ordering method based on the user and the product portrait and the association degree of the user and the product portrait according to claim 1 is characterized in that: in S1, the comprehensive commodity characteristics of the user are generated in the original user portrait by using a traditional recommendation system.
3. The search recommendation ordering method based on the user and the product portrait and the association degree of the user and the product portrait according to claim 2 is characterized in that: the traditional recommendation system is an item-based recommendation model or a user-based recommendation model.
4. The search recommendation ordering method based on the user and the product portrait and the association degree of the user and the product portrait according to claim 1 is characterized in that: and S2, generating hidden user characteristics of the commodity in the original product portrait by using a user behavior matrix algorithm in the collaborative filtering recommendation system.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911242968.4A CN111026966B (en) | 2019-12-06 | 2019-12-06 | Search recommendation ordering method based on user and product portrait and association degree of user and product portrait |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911242968.4A CN111026966B (en) | 2019-12-06 | 2019-12-06 | Search recommendation ordering method based on user and product portrait and association degree of user and product portrait |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111026966A CN111026966A (en) | 2020-04-17 |
| CN111026966B true CN111026966B (en) | 2023-12-22 |
Family
ID=70207466
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911242968.4A Active CN111026966B (en) | 2019-12-06 | 2019-12-06 | Search recommendation ordering method based on user and product portrait and association degree of user and product portrait |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111026966B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112418956A (en) * | 2020-12-16 | 2021-02-26 | 国网雄安金融科技集团有限公司 | Financial product recommendation method and device |
| CN114996589B (en) * | 2022-08-02 | 2022-10-21 | 八爪鱼人工智能科技(常熟)有限公司 | Online information pushing method and system based on epidemic prevention big data |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2000079453A2 (en) * | 1999-06-18 | 2000-12-28 | I2 Technologies, Inc. | Product substitution search method |
| JP2010092432A (en) * | 2008-10-10 | 2010-04-22 | Nec Corp | Data similarity calculation system, method, and program |
| JP2012027757A (en) * | 2010-07-26 | 2012-02-09 | Dainippon Printing Co Ltd | Commodity recommendation device, commodity recommendation system, commodity recommendation method, and program |
| CN103093376A (en) * | 2013-01-16 | 2013-05-08 | 北京邮电大学 | Clustering collaborative filtering recommendation system based on singular value decomposition algorithm |
| CN106327227A (en) * | 2015-06-19 | 2017-01-11 | 北京航天在线网络科技有限公司 | Information recommendation system and information recommendation method |
| CN107301592A (en) * | 2017-07-19 | 2017-10-27 | 北京京东尚科信息技术有限公司 | The method and device excavated for commodity substitute |
| CN108647996A (en) * | 2018-04-11 | 2018-10-12 | 中山大学 | A kind of personalized recommendation method and system based on Spark |
| CN109191240A (en) * | 2018-08-14 | 2019-01-11 | 北京九狐时代智能科技有限公司 | A kind of method and apparatus carrying out commercial product recommending |
| CN109783738A (en) * | 2019-01-22 | 2019-05-21 | 东华大学 | A kind of double extreme learning machine mixing collaborative filtering recommending methods based on more similarities |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150379609A1 (en) * | 2014-06-30 | 2015-12-31 | Kobo Incorporated | Generating recommendations for unfamiliar users by utilizing social side information |
| US11748978B2 (en) * | 2016-10-16 | 2023-09-05 | Ebay Inc. | Intelligent online personal assistant with offline visual search database |
-
2019
- 2019-12-06 CN CN201911242968.4A patent/CN111026966B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2000079453A2 (en) * | 1999-06-18 | 2000-12-28 | I2 Technologies, Inc. | Product substitution search method |
| JP2010092432A (en) * | 2008-10-10 | 2010-04-22 | Nec Corp | Data similarity calculation system, method, and program |
| JP2012027757A (en) * | 2010-07-26 | 2012-02-09 | Dainippon Printing Co Ltd | Commodity recommendation device, commodity recommendation system, commodity recommendation method, and program |
| CN103093376A (en) * | 2013-01-16 | 2013-05-08 | 北京邮电大学 | Clustering collaborative filtering recommendation system based on singular value decomposition algorithm |
| CN106327227A (en) * | 2015-06-19 | 2017-01-11 | 北京航天在线网络科技有限公司 | Information recommendation system and information recommendation method |
| CN107301592A (en) * | 2017-07-19 | 2017-10-27 | 北京京东尚科信息技术有限公司 | The method and device excavated for commodity substitute |
| CN108647996A (en) * | 2018-04-11 | 2018-10-12 | 中山大学 | A kind of personalized recommendation method and system based on Spark |
| CN109191240A (en) * | 2018-08-14 | 2019-01-11 | 北京九狐时代智能科技有限公司 | A kind of method and apparatus carrying out commercial product recommending |
| CN109783738A (en) * | 2019-01-22 | 2019-05-21 | 东华大学 | A kind of double extreme learning machine mixing collaborative filtering recommending methods based on more similarities |
Non-Patent Citations (3)
| Title |
|---|
| LevenshteinOWA operators for gene product similarity, clustering, and knowledge discovery;J.M. Keller等;《NAFIPS 2005 - 2005 Annual Meeting of the North American Fuzzy Information Processing Society》;233-234 * |
| 基于产品属性及用户偏好的个性化产品推荐方法;鲁辉等;《贵州师范大学学报》;第35卷(第1期);87-92 * |
| 隐式反馈场景中结合信任与相似度的排序推荐;廖列法等;《计算机应用研究》;第35卷(第12期);3614-3617、3631 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111026966A (en) | 2020-04-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Chang et al. | Crowd-based personalized natural language explanations for recommendations | |
| JP6543986B2 (en) | INFORMATION PROCESSING APPARATUS, INFORMATION PROCESSING METHOD, AND PROGRAM | |
| CN108460619B (en) | Method for providing collaborative recommendation model fusing explicit and implicit feedback | |
| TWI505212B (en) | A method and system for bidding on advertisements | |
| US8380727B2 (en) | Information processing device and method, program, and recording medium | |
| Cheng et al. | Video ecommerce++: Toward large scale online video advertising | |
| Li et al. | Grocery shopping recommendations based on basket-sensitive random walk | |
| CN108829847B (en) | Multi-modal modeling method based on translation and application thereof in commodity retrieval | |
| WO2021148011A1 (en) | Video processing method and device, electronic apparatus, and storage medium | |
| CN111026966B (en) | Search recommendation ordering method based on user and product portrait and association degree of user and product portrait | |
| US9619705B1 (en) | Object identification in visual media | |
| CN108153792A (en) | A kind of data processing method and relevant apparatus | |
| Zhao et al. | Mining affective words to capture customer’s affective response to apparel products | |
| Lin et al. | A consumer review-driven recommender service for web e-commerce | |
| Ahamed et al. | A recommender system based on deep neural network and matrix factorization for collaborative filtering | |
| Hidasi et al. | Enhancing matrix factorization through initialization for implicit feedback databases | |
| Goncalves et al. | The importance of brand affinity in luxury fashion recommendations | |
| JP4886940B2 (en) | MATCHING SYSTEM USING FACTOR DATABASE, MATCHING DEVICE, AND FACTOR DATABASE CREATION METHOD FOR THE SYSTEM | |
| CN116109368A (en) | Image and video instance association for e-commerce applications | |
| Knotzer | Product recommendations in e-commerce retailing applications | |
| Bao et al. | Collaborative filtering with user ratings and tags | |
| Sanders et al. | Testing the application of warranting theory to online third party marketplaces: the effects of information uniqueness and product type | |
| Padti et al. | Hybrid Movie Recommender System | |
| Xia et al. | Packaging design analysis by predicting user preference and semantic attribute | |
| CN112214641A (en) | Commodity cluster title generation method and device, computer system and readable storage medium |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |