Social data-based product matching evaluation method and system

Abstract

The invention is suitable for the technical field of electronic commerce, and provides a social data-based product matching evaluation method and system, and the method comprises the following steps: receiving registration information, verifying the registration information, and distributing a social evaluation circle for a personal account based on an interested field and an interested style after the verification is passed; receiving product matching release information, wherein the product matching release information comprises a product matching image, a field label and a style label; publishing the product matching image to a corresponding social evaluation circle according to the field label and the style label, wherein a personal account in the social evaluation circle can score and comment the product matching image; and obtaining the comprehensive score and keyword information of the product matching image according to the letter degree, score and comment of the personal account. According to the method, the social interaction and speciality between the social users are utilized, the accurate matching evaluation condition can be obtained based on a large amount of user data, and valuable product matching suggestions are provided for the users.

Description

Technical Field

[0001] The present invention relates to the technical field of e-commerce, and in particular to a product collocation evaluation method and system based on social data. Background Art

[0002] In the digital age, with the rapid development of e-commerce and social media, consumers are increasingly seeking personalized and diverse product pairing recommendations when purchasing products. Current shopping methods often rely on personal experience, sales consultants' advice, or reviews from others. This makes it difficult to obtain accurate pairing evaluation information based on more data, and the evaluation feedback is incomplete. Therefore, a method and system for product pairing evaluation based on social data is needed to address these issues. Summary of the Invention

[0003] In view of the shortcomings of the existing technology, the purpose of the present invention is to provide a product matching evaluation method and system based on social data to solve the problems existing in the above background technology.

[0004] The present invention is implemented as follows: a product collocation evaluation method based on social data, the method comprising the following steps:

[0005] Receiving registration information, the registration information including personal account, field of interest, and genre of interest;

[0006] Verify the registration information. Once verified, assign social evaluation circles to personal accounts based on areas of interest and styles of interest.

[0007] receiving product collocation release information, wherein the product collocation release information includes a product collocation image, a domain tag, and a style tag;

[0008] Publish product matching images to corresponding social evaluation circles based on domain and style tags, and personal accounts in the social evaluation circles can rate and comment on the product matching images;

[0009] Get the comprehensive score and keyword information of product matching images based on the credibility, ratings and comments of personal accounts;

[0010] Receive information on the adoption of matching items, and update the credibility of the personal account based on the information on the adoption of matching items.

[0011] As a further solution of the present invention: the step of obtaining the comprehensive score and keyword information of the product matching image based on the credibility, ratings and comments of the personal account specifically includes:

[0012] Retrieve the credibility of each personal account that rates product pairing images. Newly registered accounts will have an initial credibility score.

[0013] Calculate individual scores: individual score = convincingness × rating, and get the comprehensive score based on the average of all individual scores;

[0014] Get keyword information of product matching images based on all reviews.

[0015] As a further solution of the present invention: the step of obtaining keyword information of product matching images based on all reviews specifically includes:

[0016] Pre-process all comments based on the useless character library to remove useless characters in the comments;

[0017] Use NLP tools to segment all comments, obtain word sequences, and calculate the term frequency (TF) and inverse document frequency (IDF) of each word;

[0018] The importance of each word is determined based on the term frequency TF and the inverse document frequency IDF, and several keywords are determined according to the importance to obtain keyword information.

[0019] As a further solution of the present invention: the step of updating the credibility of the personal account based on the collocation adoption information specifically includes:

[0020] Determine the pairing effect tendency of the personal account based on the rating and the score tendency information of the personal account. The pairing effect tendency is poor or excellent. The score tendency information includes a split score. If the split score is below, the effect tendency is poor; otherwise, the effect tendency is excellent.

[0021] The propensity value K is determined based on the score and the segmentation score, and the sign of the propensity value K is determined based on the collocation adoption information and the collocation effect tendency;

[0022] The conviction base of the personal account is updated according to the propensity value K and its positive or negative sign, and the conviction degree is updated according to the conviction base, 0<conviction degree≤1.

[0023] As a further solution of the present invention: a newly registered account is provided with initial score tendency information, and the score tendency information supports user-defined modification; a newly registered account is provided with an initial conviction base, and each conviction degree corresponds to a conviction base range.

[0024] As a further solution of the present invention: the step of determining the propensity value K based on the score and the segmentation score specifically includes:

[0025] The score limit is determined based on the matching effect tendency, with a poor effect tendency corresponding to the lower score limit and an excellent effect tendency corresponding to the upper score limit;

[0026] Determine the propensity value K based on the score A, the score limit G and the split score M. , R is the constant coefficient.

[0027] Another object of the present invention is to provide a product collocation evaluation system based on social data, the system comprising:

[0028] A personal account registration module is used to receive registration information, wherein the registration information includes a personal account, fields of interest, and genres of interest;

[0029] The social evaluation circle assignment module is used to verify the registration information. After passing the verification, the social evaluation circle is assigned to the personal account based on the areas of interest and styles of interest;

[0030] A product collocation publishing module, configured to receive product collocation publishing information, wherein the product collocation publishing information includes a product collocation image, a domain tag, and a style tag;

[0031] The matching rating and comment module is used to publish product matching images to corresponding social evaluation circles based on domain tags and style tags. Individual accounts in the social evaluation circle can rate and comment on the product matching images.

[0032] A comprehensive score determination module is used to obtain the comprehensive score and keyword information of the product matching image based on the credibility, ratings and comments of the personal account;

[0033] The credibility update module is used to receive the collocation adoption information and update the credibility of the personal account based on the collocation adoption information.

[0034] As a further solution of the present invention: the comprehensive score determination module includes:

[0035] The credibility retrieval unit is used to retrieve the credibility of each personal account that rates the product matching image. Newly registered accounts are assigned an initial credibility.

[0036] Comprehensive score calculation unit, used to calculate individual scores, individual score = convincingness × score, the comprehensive score is obtained based on the average of all individual scores;

[0037] The keyword information determining unit is used to obtain keyword information of the product matching image based on all the comments.

[0038] As a further solution of the present invention: the keyword information determination unit includes:

[0039] The comment preprocessing subunit is used to preprocess all comments based on the useless character library and remove useless characters in the comments;

[0040] The word sequence processing subunit is used to segment all comments using NLP tools to obtain word sequences and calculate the term frequency (TF) and inverse document frequency (IDF) of each word.

[0041] The keyword information subunit is used to determine the importance of each word based on the term frequency TF and the inverse document frequency IDF, determine several keywords according to the importance, and obtain keyword information.

[0042] As a further solution of the present invention: the credibility update module includes:

[0043] A pairing effect tendency unit is used to determine the pairing effect tendency of the personal account based on the score and the score tendency information of the personal account. The pairing effect tendency is poor or excellent. The score tendency information includes a split score. If the split score is below, the effect tendency is poor. Otherwise, the effect tendency is excellent.

[0044] A propensity value calculation unit, configured to determine a propensity value K based on the score and the segmentation score, and determine the sign of the propensity value K based on the collocation adoption information and the collocation effect tendency;

[0045] The convincing degree updating unit is used to update the convincing degree of the personal account according to the tendency value K and its positive or negative sign, and to update the convincing degree according to the convincing degree, 0<convincing degree≤1.

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] This invention assigns social comment circles to individual accounts based on their areas of interest and styles. Product pairing images are then posted to corresponding social comment circles based on these area and style tags. Individual accounts in these social comment circles can then rate and comment on these product pairing images. This approach leverages the social interactions and expertise of social users, enabling accurate pairing assessments based on extensive user data and providing valuable product pairing recommendations. Furthermore, the credibility of individual accounts is updated based on pairing adoption information, providing users with even higher-quality product pairing assessments. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 A flowchart of a product collocation evaluation method based on social data.

[0049] Figure 2 A flowchart for obtaining a comprehensive score in a product collocation evaluation method based on social data.

[0050] Figure 3 A flowchart for obtaining keyword information in a product pairing evaluation method based on social data.

[0051] Figure 4 A flowchart for updating the credibility of a product pairing evaluation method based on social data.

[0052] Figure 5 Flowchart for determining propensity scores in a product pairing evaluation method based on social data.

[0053] Figure 6 This is a structural diagram of a product matching evaluation system based on social data. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solutions and advantages of the present invention clearer, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0055] The specific implementation of the present invention is described in detail below with reference to specific embodiments.

[0056] like Figure 1 As shown, an embodiment of the present invention provides a product collocation evaluation method based on social data, the method comprising the following steps:

[0057] S100, receiving registration information, the registration information including personal account, field of interest and genre of interest;

[0058] S200, verifying the registration information. After passing the verification, assigning a social evaluation circle to the personal account based on the areas of interest and styles of interest;

[0059] S300, receiving product collocation release information, wherein the product collocation release information includes a product collocation image, a domain tag, and a style tag;

[0060] S400, publishing the product combination image to the corresponding social evaluation circle according to the domain label and the style label, and the personal accounts in the social evaluation circle can rate and comment on the product combination image;

[0061] S500, obtains the comprehensive score and keyword information of product matching images based on the credibility, ratings and comments of personal accounts;

[0062] S600: Receive information on the adoption of the matching items, and update the credibility of the personal account based on the information on the adoption of the matching items.

[0063] It's important to note that existing product pairing evaluation methods are mostly based on algorithmic recommendations or simple user reviews, often overlooking the social connections between users and their expertise and influence in specific fields and styles. This approach is limited in that it struggles to accurately capture users' individual needs and preferences, and it also fails to fully leverage the rich user data and interactive behaviors in social networks to optimize recommendation results. The present invention aims to address these issues.

[0064] In an embodiment of the present invention, full use is made of the rich user data and interactive behaviors in the social platform to obtain accurate matching evaluation information. First, the user needs to register a social platform account by entering registration information. The registration information includes a personal account, areas of interest and styles of interest. Both areas of interest and styles of interest can contain multiple areas. Then, the embodiment of the present invention will automatically assign social evaluation circles to personal accounts based on areas of interest and styles of interest. For example, the social evaluation circles assigned to the user's personal account are: simple style clothing circle and modern style home circle. In this way, through the personal account, the user can see the content posted by other users in the simple style clothing circle and the modern style home circle, and can rate and comment on the content posted by other users. When a user wants to publish content, he or she only needs to enter the product combination publishing information, which includes product combination images, domain tags and style tags. Then, the embodiment of the present invention will publish the product combination images to the corresponding social evaluation circle according to the domain tags and style tags. All personal accounts in the social evaluation circle can rate and comment on the product combination images. In this way, the product combination images can be accurately published, because users who see the product combination images are interested in them. After the publication, users in the social evaluation circle will successively rate and comment on them. Then, at each set time interval, the embodiment of the present invention will automatically obtain the comprehensive score and keyword information of the product combination image based on the credibility, rating and comments of the personal account. The credibility of the personal account represents the influence and rating accuracy of the user. In this way, the embodiment of the present invention makes full use of the social interaction and professionalism between social users, and can obtain accurate matching evaluation based on a large amount of user data, providing users with valuable product matching suggestions. In addition, after the user makes a product matching decision, he or she needs to input matching adoption information, which is whether the matching is adopted or not adopted. Matching adoption means that the products in the product matching image are used for matching, and matching not adopted means that the products in the product matching image are not used for matching. Finally, the embodiment of the present invention will also update the credibility of the personal account based on the matching adoption information, so as to provide users with higher-quality product matching evaluation information.

[0065] like Figure 2 As shown, as a preferred embodiment of the present invention, the step of obtaining the comprehensive score and keyword information of the product matching image based on the credibility, rating and comments of the personal account specifically includes:

[0066] S501, retrieve the credibility of each personal account that rates the product matching image. A newly registered account has an initial credibility.

[0067] S502, calculate individual scores, individual score = convincingness × score, and obtain a comprehensive score based on the average of all individual scores;

[0068] S503: Obtain keyword information of product matching images based on all comments.

[0069] In an embodiment of the present invention, each new user is assigned an initial credibility after registering an account. The initial credibility is a fixed value set by the social platform. Before calculating the comprehensive score, it is necessary to retrieve the credibility of each personal account that has rated the product matching image, and then calculate the personal score. Personal score = credibility × score. Then, the comprehensive score can be obtained based on the average of all personal scores. Then, the keyword information of the product matching image can be obtained based on all the comments.

[0070] like Figure 3 As shown, as a preferred embodiment of the present invention, the step of obtaining keyword information of product matching images based on all reviews specifically includes:

[0071] S5031, pre-processing all comments based on the useless character library to remove useless characters in the comments;

[0072] S5032: Use NLP tools to segment all comments into words, obtain word sequences, and calculate the term frequency (TF) and inverse document frequency (IDF) of each word.

[0073] S5033: Determine the importance of each word based on the term frequency TF and the inverse document frequency IDF, determine several keywords based on the importance, and obtain keyword information.

[0074] In an embodiment of the present invention, in order to obtain valuable keyword information, all comments are first preprocessed based on a useless character library to remove useless characters in the comments. The useless character library needs to be established in advance. For example, the useless character library contains HTML tags, special symbols, stop words, etc. Then, all comments are segmented through NLP tools to obtain word sequences, and the word frequency TF and inverse document frequency IDF of each word are calculated. Assuming that there is a comment set D={d1, d2, d3, .., dn}, where di represents a comment, and t is the word for which the TF value is to be calculated, then TF=the number of times word t appears in di / the total number of words in di, IDF=log(the total number of comments in the comment set D / the number of comments containing word t), and finally the importance of each word is determined based on the word frequency TF and the inverse document frequency IDF, importance=k1*TF+k2*IDF, k1 and k2 are both constant coefficients, and the top few importance rankings are determined as keywords, and several keywords constitute keyword information.

[0075] like Figure 4As shown, as a preferred embodiment of the present invention, the step of updating the credibility of the personal account based on the collocation adoption information specifically includes:

[0076] S601: Determine the pairing effect tendency of the personal account based on the score and the score tendency information of the personal account. The pairing effect tendency is divided into poor effect tendency and excellent effect tendency. The score tendency information includes a split score. A score below the split score indicates a poor effect tendency, and a score above the split score indicates an excellent effect tendency.

[0077] S602, determining a propensity value K based on the score and the segmentation score, and determining the sign of the propensity value K based on the collocation adoption information and the collocation effect tendency;

[0078] S603, updating the conviction base of the personal account according to the tendency value K and its sign, and updating the conviction degree according to the conviction base, 0<conviction degree≤1.

[0079] In an embodiment of the present invention, a newly registered account is provided with initial score tendency information, and the score tendency information includes a split score. A score below the split score indicates a poor effect tendency, and a score greater than or equal to the split score indicates an excellent effect tendency. For example, the initial split score is 6 points. The score tendency information supports users to customize and modify it according to their own standards, such as changing the split score to 7, and the newly registered account is provided with an initial conviction base. Each conviction level corresponds to a conviction base range, 0<conviction ≤1. When updating the credibility, you first need to determine the matching effect tendency of the personal account based on the score and the score tendency information of the personal account. The matching effect tendency is poor effect tendency or excellent effect tendency. Then, determine the tendency value K based on the score and the split score. Determine the positive or negative sign of the tendency value K based on the matching adoption information and the matching effect tendency. For example, when the effect tendency is poor and the matching is adopted, the corresponding K value is negative; when the effect tendency is excellent and the matching is adopted, the corresponding K value is positive. Finally, update the credibility base of the personal account based on the tendency value K and its positive or negative sign. Simply add or subtract the K value from the credibility base, and update the credibility based on the credibility base.

[0080] like Figure 5 As shown in FIG. 1 , as a preferred embodiment of the present invention, the step of determining the propensity value K based on the score and the segmentation score specifically includes:

[0081] S6021: Determine the score limit based on the matching effect tendency. A poor effect tendency corresponds to the lower score limit, and an excellent effect tendency corresponds to the upper score limit.

[0082] S6022: Determine the propensity score K based on the score A, the score limit G, and the split score M. , R is the constant coefficient.

[0083] In the embodiment of the present invention, in order to determine the propensity value K, it is necessary to determine the score limit value according to the collocation effect tendency. The poor effect tendency corresponds to the score lower limit value, and the excellent effect tendency corresponds to the score upper limit value. For example, if the score range is 0-10 points, the score limit value corresponding to the poor effect tendency is 0 points, and the score limit value corresponding to the excellent effect tendency is 10 points. Then, the propensity value K can be calculated based on the score A, the score limit value G, and the split score M. , R is a fixed value set in advance. The larger K is, the stronger the user's effect tendency is.

[0084] like Figure 6 As shown, an embodiment of the present invention further provides a product collocation evaluation system based on social data, the system comprising:

[0085] A personal account registration module 100 is configured to receive registration information, including a personal account, areas of interest, and genres of interest;

[0086] The social evaluation circle assignment module 200 is used to verify the registration information and, after passing the verification, assign a social evaluation circle to the personal account based on the areas of interest and the styles of interest;

[0087] A product collocation publishing module 300 is configured to receive product collocation publishing information, wherein the product collocation publishing information includes a product collocation image, a domain tag, and a style tag;

[0088] The matching rating and commenting module 400 is used to publish product matching images to corresponding social evaluation circles based on domain tags and style tags. Individual accounts in the social evaluation circles can rate and comment on the product matching images.

[0089] Comprehensive score determination module 500, for obtaining a comprehensive score and keyword information of a product pairing image based on the credibility, ratings, and comments of a personal account;

[0090] The credibility updating module 600 is configured to receive information on the adoption of a combination and update the credibility of the personal account based on the information on the adoption of the combination.

[0091] As a preferred embodiment of the present invention, the comprehensive score determination module 500 includes:

[0092] The credibility retrieval unit is used to retrieve the credibility of each personal account that rates the product matching image. Newly registered accounts are assigned an initial credibility.

[0093] Comprehensive score calculation unit, used to calculate individual scores, individual score = convincingness × score, the comprehensive score is obtained based on the average of all individual scores;

[0094] The keyword information determining unit is used to obtain keyword information of the product matching image based on all the comments.

[0095] As a preferred embodiment of the present invention, the keyword information determination unit includes:

[0096] The comment preprocessing subunit is used to preprocess all comments based on the useless character library and remove useless characters in the comments;

[0097] The word sequence processing subunit is used to segment all comments using NLP tools to obtain word sequences and calculate the term frequency (TF) and inverse document frequency (IDF) of each word.

[0098] The keyword information subunit is used to determine the importance of each word based on the term frequency TF and the inverse document frequency IDF, determine several keywords according to the importance, and obtain keyword information.

[0099] As a preferred embodiment of the present invention, the credibility updating module 600 includes:

[0100] A pairing effect tendency unit is used to determine the pairing effect tendency of the personal account based on the score and the score tendency information of the personal account. The pairing effect tendency is poor or excellent. The score tendency information includes a split score. If the split score is below, the effect tendency is poor. Otherwise, the effect tendency is excellent.

[0101] A propensity value calculation unit, configured to determine a propensity value K based on the score and the segmentation score, and determine the sign of the propensity value K based on the collocation adoption information and the collocation effect tendency;

[0102] The convincing degree updating unit is used to update the convincing degree of the personal account according to the tendency value K and its positive or negative sign, and to update the convincing degree according to the convincing degree, 0<convincing degree≤1.

[0103] The above is only a detailed description of the preferred embodiments of the present invention, which is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

[0104] It should be understood that, although the various steps in the flow chart of each embodiment of the present invention are shown in sequence according to the indication of the arrows, these steps are not necessarily performed in sequence according to the order indicated by the arrows. Unless otherwise specified herein, the execution of these steps is not strictly limited in order, and these steps can be performed in other orders. Moreover, at least a portion of the steps in each embodiment may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.

[0105] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing the relevant hardware through a computer program. The program can be stored in a non-volatile computer-readable storage medium. When executed, the program can include the processes of the above-described method embodiments. Any reference to memory, storage, database, or other media used in the various embodiments provided herein may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct RAMbus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM).

[0106] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the disclosure in the specification and examples. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered merely as exemplary, and the true scope and spirit of the present disclosure are indicated by the claims.

Claims

1. A product collocation evaluation method based on social data, characterized in that: The method comprises the following steps: Receiving registration information, the registration information including personal account, field of interest, and genre of interest; Verify the registration information. Once verified, assign social evaluation circles to personal accounts based on areas of interest and styles of interest. receiving product collocation release information, wherein the product collocation release information includes a product collocation image, a domain tag, and a style tag; Publish product matching images to corresponding social evaluation circles based on domain and style tags, and personal accounts in the social evaluation circles can rate and comment on the product matching images; Get the comprehensive score and keyword information of product matching images based on the credibility, ratings and comments of personal accounts; Receive information on the adoption of matching items, and update the credibility of the personal account based on the information on the adoption of matching items.

2. The product collocation evaluation method based on social data according to claim 1, characterized in that: The step of obtaining the comprehensive score and keyword information of the product matching image based on the credibility, ratings and comments of the personal account specifically includes: Retrieve the credibility of each personal account that rates product pairing images. Newly registered accounts will have an initial credibility score. Calculate individual scores: individual score = convincingness × rating, and get the comprehensive score based on the average of all individual scores; Get keyword information of product matching images based on all reviews.

3. The product collocation evaluation method based on social data according to claim 2, characterized in that: The step of obtaining keyword information of product matching images based on all reviews specifically includes: Pre-process all comments based on the useless character library to remove useless characters in the comments; Use NLP tools to segment all comments, obtain word sequences, and calculate the term frequency (TF) and inverse document frequency (IDF) of each word; The importance of each word is determined based on the term frequency TF and the inverse document frequency IDF, and several keywords are determined according to the importance to obtain keyword information.

4. The product collocation evaluation method based on social data according to claim 1, characterized in that: The step of updating the credibility of the personal account based on the pairing adoption information specifically includes: Determine the pairing effect tendency of the personal account based on the rating and the score tendency information of the personal account. The pairing effect tendency is poor or excellent. The score tendency information includes a split score. If the split score is below, the effect tendency is poor; otherwise, the effect tendency is excellent. The propensity value K is determined based on the score and the segmentation score, and the sign of the propensity value K is determined based on the collocation adoption information and the collocation effect tendency; The conviction base of the personal account is updated according to the propensity value K and its positive or negative sign, and the conviction degree is updated according to the conviction base, 0<conviction degree≤1.

5. The product collocation evaluation method based on social data according to claim 4, characterized in that: Newly registered accounts are set with initial score tendency information, which can be customized by users. Newly registered accounts are set with an initial conviction base, and each conviction level corresponds to a conviction base range.

6. The product collocation evaluation method based on social data according to claim 4, characterized in that: The step of determining the propensity value K based on the score and the segmentation score specifically includes: The score limit is determined based on the matching effect tendency, with a poor effect tendency corresponding to the lower score limit and an excellent effect tendency corresponding to the upper score limit; Determine the propensity value K based on the score A, the score limit G and the split score M. , R is the constant coefficient.

7. A product matching evaluation system based on social data, characterized in that: The system comprises: A personal account registration module is used to receive registration information, wherein the registration information includes a personal account, fields of interest, and genres of interest; The social evaluation circle assignment module is used to verify the registration information. After passing the verification, the social evaluation circle is assigned to the personal account based on the areas of interest and styles of interest; A product collocation publishing module, configured to receive product collocation publishing information, wherein the product collocation publishing information includes a product collocation image, a domain tag, and a style tag; The matching rating and comment module is used to publish product matching images to corresponding social evaluation circles based on domain tags and style tags. Individual accounts in the social evaluation circle can rate and comment on the product matching images. A comprehensive score determination module is used to obtain the comprehensive score and keyword information of the product matching image based on the credibility, ratings and comments of the personal account; The credibility update module is used to receive the collocation adoption information and update the credibility of the personal account based on the collocation adoption information.

8. The product collocation evaluation system based on social data according to claim 7, characterized in that: The comprehensive score determination module includes: The credibility retrieval unit is used to retrieve the credibility of each personal account that rates the product matching image. Newly registered accounts are assigned an initial credibility. Comprehensive score calculation unit, used to calculate individual scores, individual score = convincingness × score, the comprehensive score is obtained based on the average of all individual scores; The keyword information determining unit is used to obtain keyword information of the product matching image based on all the comments.

9. The product collocation evaluation system based on social data according to claim 8, characterized in that: The keyword information determination unit includes: The comment preprocessing subunit is used to preprocess all comments based on the useless character library and remove useless characters in the comments; The word sequence processing subunit is used to segment all comments using NLP tools to obtain word sequences and calculate the term frequency (TF) and inverse document frequency (IDF) of each word. The keyword information subunit is used to determine the importance of each word based on the term frequency TF and the inverse document frequency IDF, determine several keywords according to the importance, and obtain keyword information.

10. The product collocation evaluation system based on social data according to claim 7, characterized in that: The convincingness updating module includes: A pairing effect tendency unit is used to determine the pairing effect tendency of the personal account based on the score and the score tendency information of the personal account. The pairing effect tendency is poor or excellent. The score tendency information includes a split score. If the split score is below, the effect tendency is poor. Otherwise, the effect tendency is excellent. A propensity value calculation unit, configured to determine a propensity value K based on the score and the segmentation score, and determine the sign of the propensity value K based on the collocation adoption information and the collocation effect tendency; The convincing degree updating unit is used to update the convincing degree of the personal account according to the tendency value K and its positive or negative sign, and to update the convincing degree according to the convincing degree, 0<convincing degree≤1.

Images