An intelligent management system based on a WeChat e-commerce ecological platform

Abstract

This invention relates to the field of e-commerce operation management, specifically an intelligent management system based on the WeChat e-commerce ecosystem platform. The invention categorizes stores by their main product categories, collects product information and customer profiles for each type of store, calculates the weighted similarity of common products, prices, and customer structure among stores of the same type to obtain store correlation, constructs a similarity correlation table centered on each store, and imports it into the platform. It analyzes the purchase impact of each feature video by segmenting videos on video detail pages and combining this with order video playback data. It calculates the interest impact using completion rate and interest factors, and then combines this with the purchase impact to obtain a comprehensive interest score. Finally, it dynamically sorts the videos based on the comprehensive interest scores of each store's feature videos. This system provides the platform with accurate correlations between stores, improving the efficiency of collaborative store operations and the conversion rate of product video marketing on the WeChat e-commerce platform.

Description

Technical Field

[0001] This invention relates to the field of e-commerce operation management, and specifically to an intelligent management system based on the WeChat e-commerce ecosystem platform. Background Technology

[0002] In the current WeChat e-commerce ecosystem, the number and types of stores on the platform have surged, and product information and user data are growing exponentially. On the one hand, the platform struggles to accurately grasp the relationships between stores and cannot make precise recommendations based on user needs and experience. On the other hand, as the core marketing medium, the video content on product detail pages lacks scientific quantitative assessment of the impact of different playback orders on users' purchasing decisions, resulting in low conversion rates for video marketing. Therefore, there is an urgent need for an intelligent management system that can integrate store data, quantify video value, and dynamically optimize marketing strategies.

[0003] However, existing technologies have the following problems: 1. Existing technologies mostly only establish simple associations around a single product dimension, without integrating product structure, pricing strategy and customer characteristics at the overall store level to build a systematic association analysis system; this makes it difficult for the platform to optimize the overall merchant layout and provide merchants with accurate operational directions, ultimately resulting in scattered store resources, low traffic utilization efficiency, and difficulty for merchants to achieve growth through ecosystem collaboration.

[0004] 2. Existing video displays mostly adopt fixed sorting or manual adjustment modes, without considering the impact of video content playback order on marketing effectiveness, without establishing a system based on multi-dimensional data analysis of the impact of video content on user behavior, and without establishing a dynamic optimization mechanism; resulting in video display strategies being out of touch with user needs and market changes, and failing to continuously guarantee marketing effectiveness. Summary of the Invention

[0005] This invention aims to address the shortcomings of existing technologies by providing an intelligent management system based on the WeChat e-commerce ecosystem platform. The system first categorizes and analyzes stores to construct a store similarity association table, which is then uploaded to the WeChat e-commerce ecosystem platform. Next, it quantitatively evaluates product-feature videos from two dimensions: purchase impact and user interest. Finally, based on the evaluation results, it dynamically optimizes video display strategies to help merchants improve marketing conversion efficiency, while providing platform users with more tailored content recommendations.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: an intelligent management system based on the WeChat e-commerce ecosystem platform, comprising a store segmentation and information collection module, a store association analysis module, a video analysis module, a video content comprehensive evaluation module, and a marketing decision-making module. The connections between the modules are as follows: the store segmentation and information collection module is connected to the store association analysis module; the video analysis module is connected to both the store association analysis module and the video content comprehensive evaluation module; and the marketing decision-making module is connected to the video content comprehensive evaluation module.

[0007] Store Classification and Information Collection Module: Based on the main category of the store applied for when merchants apply to join the platform, the store is classified into categories, and product information and customer profiles of each store in each category are obtained from the platform database.

[0008] Store Association Analysis Module: Calculates the association degree between stores in each category by combining product information and customer profiles, establishes a similarity association table centered on each store based on the association degree, and imports the similarity association tables of each store into the platform.

[0009] Video analytics module: Retrieves user behavior data for each product from the platform database, extracts sales information for each product in each store corresponding to each order, analyzes the contribution of each feature video, and analyzes the purchase impact value of each feature video through the contribution.

[0010] The video content comprehensive evaluation module extracts the operation behavior and playback data of each user in each store from the user behavior data of each product, analyzes the interest influence of each feature video, and analyzes the comprehensive interest score of each feature video through purchase influence value and interest influence.

[0011] Marketing Decision Module: After processing the videos based on the comprehensive interest scores of each feature video, the feature videos of each store are dynamically sorted.

[0012] Compared with the prior art, the present invention has the following beneficial effects: (1) The present invention classifies stores by the main category of the store in the information uploaded by the merchant when applying to join the platform, and obtains the product information and customer profile of each store in each type from the platform database, which facilitates the subsequent correlation analysis of stores, provides core data source for multiple modules, and ensures that the subsequent analysis can be implemented.

[0013] (2) This invention calculates the correlation between each store in each type by combining product information and customer profile, establishes a similarity correlation table centered on each store by correlation, imports the similarity correlation table of each store into the platform, provides the platform with a basis for store collaborative operation, strengthens ecological synergy, and improves overall operational efficiency and control capabilities.

[0014] (3) This invention obtains user behavior data of each product from the platform database, extracts the sales information of each product in each store corresponding to each order, analyzes the contribution of each feature video, analyzes the purchase impact value of each feature video through the contribution analysis, provides a data basis for subsequent analysis of the comprehensive interest score of each feature video, and improves the accuracy of feature video value judgment.

[0015] (4) This invention extracts the user's operation behavior and playback data from the user behavior data of each product, analyzes the interest influence of each feature video, and analyzes the comprehensive interest score of each feature video through purchase influence value and interest influence. Multi-dimensional analysis improves the accuracy of evaluation and provides a basis for subsequent processing and ranking of video marketing. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the system module connections of the present invention.

[0018] Figure 2 This is a schematic diagram illustrating the steps involved in calculating the correlation between stores in this invention.

[0019] Figure 3 This is a schematic diagram of the steps in the analysis method of interest influence degree in this invention. Detailed Implementation

[0020] Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps set forth in these embodiments do not limit the scope of the invention. Furthermore, it should be understood that, for ease of description, the dimensions of the various parts shown in the drawings are not drawn to actual scale.

[0021] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the invention or its application or use. Techniques, methods, and apparatus known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and apparatus should be considered part of the specification.

[0022] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0023] This invention relates to the field of e-commerce operation management, specifically an intelligent management system based on the WeChat e-commerce ecosystem platform. The invention categorizes stores by their main product categories, collects product information and customer profiles for each type of store, calculates the weighted similarity of common products, prices, and customer structure among stores of the same type to obtain store correlation, constructs a similarity correlation table centered on each store, and imports it into the platform. It analyzes the purchase impact of each feature video by segmenting videos on video detail pages and combining this with order video playback data. It calculates the interest impact using completion rate and interest factors, and then combines this with the purchase impact to obtain a comprehensive interest score. Finally, it dynamically sorts the videos based on the comprehensive interest scores of each store's feature videos. This system provides the platform with accurate correlations between stores, improving the efficiency of collaborative store operations and the conversion rate of product video marketing on the WeChat e-commerce platform.

[0024] Please see Figure 1 As shown, this invention provides an intelligent management system based on the WeChat e-commerce ecosystem platform, including a store segmentation and information collection module, a store association analysis module, a video analysis module, a video content comprehensive evaluation module, and a marketing decision-making module. The connections between the modules are as follows: the store segmentation and information collection module is connected to the store association analysis module; the video analysis module is connected to both the store association analysis module and the video content comprehensive evaluation module; and the marketing decision-making module is connected to the video content comprehensive evaluation module.

[0025] Store Classification and Information Collection Module: Based on the main category of the store applied for when merchants apply to join the platform, the store is classified into categories, and product information and customer profiles of each store in each category are obtained from the platform database.

[0026] It should be noted that: product information includes product model, price, and total number of products, where the total number of products refers to the number of different models; customer profile includes user gender, age, and core interest tags.

[0027] In one embodiment of the present invention, considering that there are many merchants and a wide range of business categories on the WeChat e-commerce platform, directly calculating the correlation degree would result in results without practical reference value. Therefore, each store is classified by its main category to ensure that in the subsequent store correlation analysis module, only the correlation degree of stores of the same type is calculated, so that the correlation degree results have operational guidance significance.

[0028] In addition, it should be noted that obtaining product information and customer profiles for each store in each category is for the purpose of calculating the correlation between stores later.

[0029] This invention categorizes stores based on their main product categories uploaded when merchants apply to join the platform, and retrieves product information and customer profiles for each store in each category from the platform's database. This facilitates subsequent store correlation analysis, provides a core data source for multiple modules, and ensures that subsequent analysis can be implemented.

[0030] Store Association Analysis Module: Calculates the association degree between stores in each category by combining product information and customer profiles, establishes a similarity association table centered on each store based on the association degree, and imports the similarity association tables of each store into the platform.

[0031] In a preferred example, please refer to Figure 2 As shown, the specific steps for comprehensively calculating the correlation between stores are as follows: W1, obtain the product model, price and total number of products in the product information of each store in each type, and count the number and price of products with the same model between two stores of the same type.

[0032] Considering that directly using the sum of the total number of goods may lead to a distortion of the proportion due to the large difference in the total number of goods between the two stores, W2, when the number of goods of the same model between the two stores is not zero, calculate the average of the total number of goods between the two stores, and record the ratio of the number of the same goods to the average number of goods as the common proportion of goods between the two stores.

[0033] Considering that when users compare similar stores, the strongest signal of perceived price similarity is that the prices of the same products are completely identical, and considering that the ratio can convert price differences into relative proportions, a unified quantitative standard can be established to objectively reflect the degree of price similarity between the two stores.

[0034] Based on this, W3 compares the prices of identical products between two stores of the same type. When prices are equal, the maximum price similarity is set as the price similarity; otherwise, the ratio of the lowest to the highest price is used as the price similarity. The price similarity of each pair of identical products is calculated, and the average value is taken as the price similarity between the two stores. The maximum price similarity is set to 1.

[0035] W4. Obtain the customer profile feature distribution vector of the store based on the customer profile, calculate the cosine similarity of the customer profile feature distribution vectors between the two stores, and denot it as the customer structure similarity.

[0036] It should be further explained that the specific method for calculating the cosine similarity of the customer profile feature distribution vector is as follows: extract gender, age and core interest tags from the customer profiles provided by the platform, divide gender, age and core interest tags into different feature dimensions, and count the proportion of customers in each feature dimension for each store to obtain the customer profile feature distribution vector corresponding to each store.

[0037] The cosine similarity between the customer profile feature distribution vectors of two stores is calculated using the cosine similarity formula.

[0038] In a preferred example, when dividing different feature dimensions, gender is first divided into two dimensions: male and female. Age is divided into three dimensions: under 30 years old, 31-50 years old, and over 51 years old. Core interest tags are divided into seven dimensions: clothing and fashion, digital home appliances, maternal and infant care, home life, beauty and skin care, sports and outdoor, and food and fresh produce.

[0039] The core interest tags for each user are set by calculating the total number of product categories in each list based on the user's historical purchase records, favorites list, and shopping cart list. The ratio of the number of products corresponding to each interest tag to the total number of product categories is then calculated, and the tag with the largest ratio is taken as the user's core interest tag.

[0040] W5. The correlation between two stores is calculated by weighting the proportion of common products, price similarity, and customer structure similarity.

[0041] Preferably, in a specific embodiment of the present invention, the weights for the weighted calculation can be set as follows: the weight of common product proportion is 40%, the weight of price similarity is 40%, and the weight of customer structure similarity is 20%. The implementer can also adjust the specific values ​​according to the actual situation, but it is necessary to ensure that the sum of each weight is equal to 1.

[0042] W6. Calculate the correlation between any two stores of the same type within the platform in turn, and finally obtain the correlation between each store.

[0043] It should be noted that the similarity association table is built by forming a set of stores of the same type, and then iterating through the various sets of stores of different types on the platform.

[0044] For each store in each set, take it as the central store, calculate the correlation between the central store and every other store in the set, and form a correlation array.

[0045] Using the central main store as the unique primary key, the array is sorted in descending order of its relevance value and stored sequentially to generate a similarity association table corresponding to the central main store.

[0046] The final result is a similarity association table for each store, providing core data support for subsequent precise platform operations, merchant resource collaboration, and user demand matching.

[0047] This invention calculates the correlation between stores in each category by combining product information and customer profiles. It then establishes a similarity correlation table centered on each store based on the correlation, imports the similarity correlation tables of each store into the platform, provides the platform with a basis for collaborative store operation, strengthens ecological synergy, and improves overall operational efficiency and control capabilities.

[0048] Video analytics module: Retrieves user behavior data for each product from the platform database, extracts sales information for each product in each store corresponding to each order, analyzes the contribution of each feature video, and analyzes the purchase impact value of each feature video through the contribution.

[0049] Considering that traditional stores mostly rely on fixed video content playback data, it's not possible to directly analyze which part of the video plays a core role in marketing. Furthermore, considering that video playback data of sold products in the store can reflect the value of the video, and that by converting the actual impact of the video on purchases into calculable data, a crucial basis for comprehensive evaluation can be provided.

[0050] Based on this, the specific content of the video analysis module includes: segmenting the entire introductory video of each product video detail page of each store according to the video content to obtain videos with various features.

[0051] It should be added that the video segmentation method is as follows: based on traditional computer vision technology, the video is segmented according to different video content themes. For example, a video introducing a mobile phone can be segmented into videos with different features according to the camera technology, earpiece, appearance, performance and mobile phone system.

[0052] Extract the playback records and contribution views of user-specific videos from the sales information of each product in each store and each order.

[0053] It should be noted that the contribution play count refers to the proportion of the number of times each feature video is played in a single purchase record to the sum of the number of times each feature video is played.

[0054] The contribution of each feature video is calculated based on its playback history and the number of views it contributes.

[0055] In one embodiment of the present invention, the specific method for comprehensively calculating the contribution of each feature video is to determine the playback position number of each feature video based on the playback record.

[0056] The total number of feature videos on the product video detail page for each order and the playback position number of each feature video are obtained. The weight coefficient of each feature video is calculated by weighted linear attribution.

[0057] The formula for calculating the weighted linear attribution is as follows: .

[0058] in Representing the The weight coefficients of each feature video. Representing the The playback sequence number of each characteristic video. The total number of feature videos on the product video detail page. The ID representing the characteristic video, where .

[0059] In the above formula, This is to ensure that the weights change monotonically with the playback sequence number as the position changes. It represents the sum of all position indices, satisfying the mathematical constraint that the sum of all weights is 1, thus ensuring the standardization and additivity of the weights.

[0060] The above formula ensures that videos with larger position numbers (i.e., those played later in the playback order) receive higher weights, and the sum of all weight coefficients is 1.

[0061] The ratio of the playback contribution of each feature video to the sum of the playback contribution of each feature video is used as the playback contribution coefficient.

[0062] The contribution of each feature video is calculated by multiplying its playback contribution coefficient and weight coefficient. This multiplication of the playback contribution coefficient and weight coefficient reflects the contribution of each feature video, achieving multi-dimensional fusion and more accurately reflecting the contribution.

[0063] The feature video with the highest contribution from the product video detail page in each order is selected as the core video. The number of orders with each feature video as the core video in each store is counted, and this number is recorded as the purchase impact value of each feature video.

[0064] This invention obtains user behavior data for each product from the platform database, extracts sales information corresponding to each order for each product in each store, analyzes the contribution of each feature video, and analyzes the purchase impact value of each feature video through contribution analysis, providing a data foundation for subsequent analysis of the comprehensive interest score of each feature video and improving the accuracy of feature video value judgment.

[0065] The video content comprehensive evaluation module extracts the operation behavior and playback data of each user in each store from the user behavior data of each product, analyzes the interest influence of each feature video, and analyzes the comprehensive interest score of each feature video through purchase influence value and interest influence.

[0066] Considering that the purchase impact value obtained solely from the video analytics module cannot fully reflect the value of characteristic videos—for example, some videos may not have led to a purchase yet, but they can significantly arouse user interest and have potential conversion potential—this is also a major characteristic for measuring the value of a video.

[0067] It also considers extracting user behavior and playback data to transform interests into calculable interest impact, providing a core basis for video value assessment based on user interest dimensions.

[0068] Please refer to this. Figure 3 The analysis methods for interest influence include: S1, extracting playback data and operation behavior of each user in each store corresponding to each product from the user behavior data of each product.

[0069] S2. Extract the viewed characteristic videos and their playback completion rates from the playback data, and obtain the total number of users who played each characteristic video from the platform database; this provides a data foundation for subsequent calculation of interest factors.

[0070] S3. Calculate the completion rate of each feature video by measuring the completion rate of each user's browsing of each feature video and the total number of users who played the video.

[0071] In one specific instance, the completion rate is calculated as the ratio of the number of users who completed more than 99% of each video playback to the total number of users who played each video.

[0072] S4. Match each user's operational behavior with the set interest behavior, obtain the characteristic videos viewed by each successfully matched user, analyze the number of users who generated interest in each characteristic video, and analyze the interest factors of each characteristic video in combination with the total number of users who played each characteristic video.

[0073] Preferably, in one embodiment of the present invention, the specific analysis method of the interest factor is as follows: matching user actions with set interest behaviors; when the user action is to favorite or add to cart, the match is successful. The set interest behaviors are: favorite and add to cart.

[0074] Users who are successfully matched are recorded as users who generate interest, and videos with various features viewed by each successfully matched user are recorded as videos that generate interest.

[0075] Count the number of users who are interested in the videos with each feature.

[0076] The ratio of the number of users who generate interest in each feature video to the total number of users who play each feature video is denoted as the interest factor of each feature video.

[0077] S5. The interest impact is calculated by multiplying the interest factor and completion rate of each feature video. By multiplying the interest factor and completion rate of each feature video, both the user's potential interest in the content and the feedback from actual viewing behavior are considered. This allows for a more comprehensive and accurate measurement of the influence of the feature video on the user's interest, and is therefore a reasonable way to calculate the interest impact.

[0078] In a preferred embodiment of the present invention, the specific method for analyzing the comprehensive interest score of each feature video is as follows: the purchase impact value of each feature video is subjected to min-max standardization, and the standardized purchase impact value is used as the purchase impact degree; through standardization, purchase impact values ​​of different magnitudes are uniformly mapped to the [0, 1] interval, eliminating numerical scale differences, and ensuring that the purchase impact degree and interest impact degree are in the same numerical interval, so as to ensure that the comprehensive interest score can truly reflect the value of the video.

[0079] The purchase impact of each feature video and the interest impact of the corresponding feature video are weighted and summed to obtain the comprehensive interest score. Then, the comprehensive interest score is multiplied by the total score of the set comprehensive interest score to obtain the comprehensive interest score of each feature video.

[0080] In a specific embodiment of the present invention, during the process of obtaining the comprehensive interest score through weighted summation, since the comprehensive interest score reflects the attractiveness of each feature video to the user and the commercial value of the video, and the purchase behavior is directly related to the completion of the transaction, it is the most direct manifestation of commercial value. Its importance is greater than behaviors such as collection and adding to the shopping cart, which only express potential interest. Therefore, the weight of the purchase influence is greater than the weight of the interest influence. In the present invention, the weight is set as follows: the weight of the purchase influence is 60%, and the weight of the interest influence is set to 40%. The implementer can adjust the weights himself, but the sum of the two must be 1, and it must be ensured that the requirement that the weight of the purchase influence is greater than the weight of the interest influence is met.

[0081] The total score for the comprehensive interest rating can be set to 100 points, and the implementer can also adjust the setting value.

[0082] This invention extracts user operation behavior and playback data from user behavior data of various products, analyzes the interest influence of each feature video, and analyzes the comprehensive interest score of each feature video by purchasing influence value and interest influence. The multi-dimensional analysis improves the accuracy of the evaluation and provides a basis for subsequent processing and ranking of video marketing.

[0083] Marketing Decision Module: After processing the videos based on the comprehensive interest scores of each feature video, the feature videos of each store are dynamically sorted.

[0084] Considering that when the overall interest score of a certain feature video is too low, it means that the video is not effective in stimulating user interest and purchasing behavior, and cannot provide effective value for product marketing, continuing to display it may occupy the details page resources and reduce the user browsing experience.

[0085] Based on this, the marketing decision module includes: comparing the comprehensive interest score of each feature video with a set score threshold, and deleting feature videos whose comprehensive interest score is lower than the set score threshold.

[0086] The remaining feature videos after deletion are sorted in descending order according to the comprehensive interest score of each feature video.

[0087] The videos with adjusted features are combined and displayed on the corresponding product video detail pages for each store. This ensures that users see the video with the highest overall interest rating first when they enter the product video detail page, thereby increasing product conversion rates and boosting store sales.

[0088] This invention removes videos with scores below a set threshold, sorts the remaining videos in descending order based on their comprehensive interest scores, and adjusts the sorting order in real time to achieve dynamic updates. This results in a video display that better meets user needs, improving user satisfaction with the platform and enhancing user retention and repurchase intentions.

[0089] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0090] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0091] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0092] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0093] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0094] Finally, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An intelligent management system based on the WeChat e-commerce ecosystem platform, characterized in that, Includes the following steps: Store Classification and Information Collection Module: Based on the main category of the store applied for when merchants apply to join the platform, the store is classified into different types, and the product information and customer profiles of each store in each type are obtained from the platform database. Store Association Analysis Module: Calculates the association degree between stores in each category by combining product information and customer profiles, establishes a similarity association table centered on each store based on the association degree, and imports the similarity association tables of each store into the platform; Video analytics module: Retrieves user behavior data for each product from the platform database, extracts sales information for each product in each store corresponding to each order, analyzes the contribution of each feature video, and analyzes the purchase impact value of each feature video through the contribution. The specific content of the video analysis module includes: The entire introductory video of each product's video detail page in each store is divided according to the video content to obtain videos with different characteristics; Extract the playback records and contribution of each user's videos for each feature from the sales information of each product in each store for each order; The contribution of each feature video is calculated by combining its playback history and the number of views it contributes. The feature video with the highest contribution from the product video detail page in each order is selected as the core video. The number of orders with each feature video as the core video in each store is counted and recorded as the purchase impact value of each feature video. Video content comprehensive evaluation module: By extracting the operation behavior and playback data of each user in each store from the user behavior data of each product, the interest influence of each feature video is analyzed, and the comprehensive interest score of each feature video is analyzed by purchasing influence value and interest influence. Marketing Decision Module: After processing the videos based on the comprehensive interest scores of each feature video, the feature videos of each store are dynamically sorted.

2. The intelligent management system based on the WeChat e-commerce ecosystem platform according to claim 1, characterized in that: The specific steps for comprehensively calculating the correlation between various stores are as follows: Retrieve product models, prices, and total quantities from product information for each store in each category; count the quantity and price of products with the same model between two stores of the same category. When the number of identical products between two stores is not zero, calculate the average of the total number of products between the two stores, and record the ratio of the number of identical products to the average number of products as the common product percentage between the two stores. Compare the prices of identical products in two stores of the same type. When the prices are equal, the maximum price similarity is set as the price similarity. Otherwise, the ratio of the lowest price to the highest price is used as the price similarity. Calculate the price similarity of each pair of identical products and take the average as the price similarity between the two stores. Based on the customer profile, obtain the customer profile feature distribution vector of the store, calculate the cosine similarity of the customer profile feature distribution vectors between the two stores, and denot it as the customer structure similarity. The correlation between two stores is calculated by weighting the proportion of common products, price similarity, and customer structure similarity. The correlation between any two stores of the same type within the platform is calculated sequentially to obtain the correlation between all stores.

3. The intelligent management system based on the WeChat e-commerce ecosystem platform according to claim 2, characterized in that: The specific method for calculating the cosine similarity of the customer profile feature distribution vectors is as follows: Extract gender, age, and core interest tags from the customer profiles provided by the platform, divide gender, age, and core interest tags into different feature dimensions, and calculate the customer distribution ratio of each store in each feature dimension to obtain the customer profile feature distribution vector corresponding to each store. The cosine similarity between the customer profile feature distribution vectors of two stores is calculated using the cosine similarity formula.

4. The intelligent management system based on the WeChat e-commerce ecosystem platform according to claim 1, wherein the similarity association table is established in the following way: Group all stores of the same type into a category store set, and iterate through each category store set in the platform. For each store in each set, take it as the central main store, calculate the correlation between the central main store and every other store in the set, and form a correlation array. Using the central main store as the unique primary key, the array is sorted in descending order of the relevance value from high to low and stored sequentially to generate a similarity association table corresponding to the central main store. Finally, a similarity association table for each store was obtained.

5. The intelligent management system based on the WeChat e-commerce ecosystem platform according to claim 1, characterized in that: The specific method for comprehensively calculating the contribution of each feature video is as follows: Determine the playback position number of each feature video based on the playback history; The total number of feature videos on the product video details page of each order and the playback position number of each feature video are obtained. The weight coefficient of each feature video is calculated by weighted linear attribution. The ratio of the playback contribution of each feature video to the sum of the playback contribution of each feature video is used as the playback contribution coefficient. The product of the playback contribution coefficient and the weight coefficient corresponding to each feature video is taken as the contribution degree of each feature video.

6. The intelligent management system based on the WeChat e-commerce ecosystem platform according to claim 1, characterized in that: The specific analysis methods for the influence of interest include: Extract playback data and operational behavior of each user in each store corresponding to each product from the user behavior data of each product; Extract the viewed videos with various characteristics and their completion rates from the playback data, and obtain the total number of users who played each video with various characteristics from the platform database; The completion rate of each feature video is calculated by the completion rate of each user's viewing of each feature video and the total number of users playing each video. The system matches each user's actions with their defined interests, obtains the characteristic videos viewed by each successfully matched user, analyzes the number of users who are interested in each characteristic video, and combines this with the total number of users who have played each characteristic video to analyze the interest factors of each characteristic video. The interest influence is calculated by multiplying the interest factor and completion rate of each feature video.

7. The intelligent management system based on the WeChat e-commerce ecosystem platform according to claim 6, characterized in that: The specific analysis method for the interest factors is as follows: Match user actions with defined interests; a successful match is achieved when a user's action is to add an item to their favorites or shopping cart. Users who are successfully matched are recorded as users who generate interest, and videos with various features viewed by each successfully matched user are recorded as videos that generate interest. Count the number of users who are interested in videos with each feature; The ratio of the number of users who generate interest in each feature video to the total number of users who play each feature video is denoted as the interest factor of each feature video.

8. The intelligent management system based on the WeChat e-commerce ecosystem platform according to claim 1, characterized in that: The specific method for analyzing the comprehensive interest score of videos with various features is as follows: The purchase impact value of each feature video is normalized by min-max, and the normalized purchase impact value is used as the purchase impact degree. The purchase impact of each feature video and the interest impact of the corresponding feature video are weighted and summed to obtain the comprehensive interest score. Then, the comprehensive interest score is multiplied by the total score of the set comprehensive interest score to obtain the comprehensive interest score of each feature video.

9. The intelligent management system based on the WeChat e-commerce ecosystem platform according to claim 1, characterized in that: The marketing decision-making module includes: The overall interest score of each feature video is compared with a set score threshold, and feature videos with an overall interest score lower than the set score threshold are deleted. The remaining feature videos after deletion are sorted in descending order according to the comprehensive interest score of each feature video; The videos with different features, after being rearranged in order, will be combined and displayed on the corresponding product video detail pages of each store.

Images