Man-machine interaction selected article supply system, method and application
Through the human-computer interaction product selection and supply system, the Item-CF recommendation algorithm and multi-dimensional array structure are used to solve the problem of users screening content of interest in massive information, and personalized recommendation services are realized, which improves user satisfaction and sales.
Patent Information
- Application Number
- CN202411699069.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-26
- Publication Date
- 2025-03-21
AI Technical Summary
The existing technology is difficult to effectively solve the problem of users screening content of interest in massive information, and traditional search and classification methods cannot meet users' personalized needs.
The product selection and supply system with human-computer interaction is adopted to obtain consumer historical product purchase information data, define a multi-dimensional array structure, calculate the similarity between products based on the Item-CF recommendation algorithm, generate a product recommendation list, and optimize the replenishment points through the optimal supply route module, and finally adjust the recommendation list through user feedback.
It has realized personalized recommendation services, improved user satisfaction and loyalty, effectively improved consumer participation and sales, and enhanced brand image and user reputation.
Smart Images
Figure CN119671667A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of Internet of Things, and in particular to a human-computer interactive product selection and supply system, method and application. Background Art
[0002] With the continuous development of the Internet and e-commerce, users are faced with a huge amount of information every day, such as goods, content, services, etc. This information makes it difficult for users to effectively filter out the content they are interested in, leading to information overload. With the diversification of user needs, traditional search and classification methods can no longer meet the personalized needs of users. Users hope to find content that suits their interests and preferences in a large amount of information. Merchants hope to analyze consumers' shopping preferences through their purchasing behavior and sell the products that consumers like on smart vending machines. Summary of the invention
[0003] In order to overcome the above-mentioned problems existing in the prior art, the present invention proposes a human-computer interactive product selection and supply system, method and application.
[0004] The technical solution adopted by the present invention to solve the technical problem is: a human-computer interactive product selection and supply method, which specifically includes the following steps: Step 1, obtain the consumer's historical purchase information data, define a multidimensional array structure, and perform data processing on the multidimensional array structure; Step 2: Calculate the similarity between products based on the Item-CF recommendation algorithm, generate a product recommendation list, and introduce new products; Step 3, through the optimal supply route module, generate the optimal route for the supply replenishment point according to the location of the system; Step 4, obtain the latest product order record, update the product information data in real time, and count the types and quantities of the products in the product recommendation list generated in step 2 in the latest product order record, and calculate the accuracy rate; Step 5: After purchasing the product, the purchase experience is rated through the user feedback module to collect consumers’ feedback on the recommended product results; Step 6: Adjust the product recommendation list based on the accuracy obtained in step 4 and the information collected in step 5.
[0005] In the above-mentioned human-computer interactive product selection and supply method, the specific method of defining the multidimensional array structure in step 1 is: in the product selection and supply system, corresponding scenes are set for devices placed in different scenes, and detailed information and purchase evaluation behaviors of the products in the devices are recorded. The product order data set module collects and displays order records of consumers purchasing products from different scenes, forming a four-dimensional array of consumer-device-product-scene.
[0006] In the above-mentioned human-computer interactive product selection and supply method, the data processing in step 1 specifically includes: clearing invalid or redundant records in the four-dimensional array data, processing missing values with mean filling, and constructing the data into a consumer-product interaction matrix, a consumer-device interaction matrix, a device-product interaction matrix, and a consumer-product interaction matrix under the scenario.
[0007] In the above-mentioned human-computer interactive product selection and supply method, step 2 specifically includes: calculating the similarity of products by using the Jaccard similarity calculation method, generating a set of similar products, calculating the average score of similar products by target users, and generating a product recommendation list based on the score.
[0008] In the above-mentioned human-computer interactive product selection and supply method, the specific method for calculating the average score of target users for similar products is: obtaining a user-product rating matrix through a user feedback module, obtaining a product similarity matrix based on a set of similar products, extracting products similar to the target product from the product similarity matrix, and setting a threshold to eliminate products with a similarity less than the threshold, calculating the target user score for each product with a similarity greater than the threshold, and calculating the average of the target user scores of all products with a similarity greater than the threshold to obtain the average score.
[0009] A human-computer interactive product selection and supply system, based on the above-mentioned human-computer interactive product selection and supply method, includes a product order data set module, a recommendation algorithm module, an optimal supply route module, and a user feedback module. The product order data set module is used to collect and display product purchase orders; the recommendation algorithm module is used to generate a product recommendation list based on historical order data in the product order data set module; the optimal supply route module is used to generate an optimal route for supply replenishment points based on equipment geology; the feedback module is used to collect user feedback on product selection and optimize the recommendation algorithm model based on the feedback.
[0010] In the above-mentioned human-computer interactive product selection and supply system, the Item-CF recommendation algorithm calculates the similarity between products, generates a recommendation list based on the similarity of the products and the historical behavior of consumers, and merchants select products for sale based on the intelligent list, collect the latest order records of consumers purchasing products, and count the products and quantities purchased from the recommendation list.
[0011] An application of the above-mentioned human-computer interactive product selection and supply system in a smart vending cabinet.
[0012] The beneficial effects of the present invention are as follows: (1) The present invention can provide consumers with personalized recommendation services based on their historical behaviors and preferences. This personalized recommendation method can make it easier for users to find items or content that meet their interests and needs, thereby improving consumer satisfaction and loyalty. (2) The present invention can effectively improve consumer participation. By providing consumers with recommended content of interest, it can stimulate consumers' enthusiasm for participation and increase consumers' stay time and number of interactions on the platform, which helps to improve the platform's activity and user stickiness. (3) The present invention effectively improves sales and conversion rates: based on consumers' purchase history and preferences, the smart cabinets in the area where the consumers are located sell goods that they may be interested in, thereby improving sales and conversion rates. (4) The present invention can enhance brand image: by providing consumers with accurate recommendation services, it can enhance the brand image and user reputation of smart cabinets, thereby attracting more consumers to purchase and partners who purchase smart cabinets. BRIEF DESCRIPTION OF THE DRAWINGS
[0013] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION
[0014] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention is described in detail below in conjunction with the accompanying drawings and specific implementation methods.
[0015] This embodiment discloses a human-computer interactive product selection and supply system applied to smart vending machines, which is mainly used to segment service scenarios for merchants holding smart vending devices. According to the consumption levels of different cities, different regions, different scenarios, different channels, and consumption needs of different groups of people, the above conditions are formed into association rules based on the system platform and recommendation algorithm, and personalized product configurations are formulated and provided. Consumers select and purchase goods on smart vending devices, and form a product order data set to the system platform. The system platform includes a product order data set module, a recommendation algorithm module, an optimal supply route module, and a user feedback module.
[0016] Product order data set module: collects and displays the purchase orders generated by consumers on all smart devices, and can filter, display and export product sales order details based on fields such as time period, product name, product type, and product brand.
[0017] Recommendation algorithm module: recommendation algorithm based on collaborative filtering. The collaborative filtering algorithm is based on the assumption that consumers who like similar products may have the same or similar tastes and preferences, including the Item-CF recommendation algorithm. Based on the historical product order data in the product order data set module, based on the Item-CF recommendation algorithm, the Item-CF algorithm calculates the similarity between products, and then generates a recommendation list for merchants on smart devices in the system based on the similarity of the products and the historical behavior of consumers. Merchants on smart devices select products from the recommendation list and put them on sale, collect the latest consumer purchase order records, and count the hit products and quantities.
[0018] Optimal supply route module: Based on the automatically reported location of the smart vending equipment as the address, the system can generate the optimal route for the supply replenishment point based on the equipment address, saving the replenisher's time in traveling to multiple equipment locations.
[0019] Product feedback module: The system pushes questionnaires to consumers based on their historical purchase orders, and can provide feedback on product selection based on consumer opinions.
[0020] The specific process is as follows: Generate a recommendation list based on association rules. Assuming that the consumer's shopping behavior before May is used as the data source, first obtain the user's purchase behavior data through SQL scripts (the product order data set module in the system platform), and then use the collaborative filtering component Item-CF algorithm to calculate the items closest to each item, so as to analyze and obtain multiple products that each consumer may buy at the same time.
[0021] Real shopping behavior after May. For example, if user A purchased product A before May, and product A is strongly related to product B, product B will be listed for sale in the smart cabinet and recommended to user A after May.
[0022] Count the number of recommendations and hits. Full table statistics-1 shows the product recommendation list generated based on shopping behavior before May, and full table statistics-2 shows the hit recommended products. This provides ideas for merchants to introduce new product selections in the future.
[0023] The main implementation methods and steps of the human-computer interactive product selection and supply system applied to intelligent vending machines are as follows: S1 system platform data entry, data collection and preprocessing 1. Collection of historical purchase order data of users (1) Equipment and scene information organization In the system, corresponding scenes, stores, addresses, equipment models, operating status and other information are set for smart vending devices deployed in different channels in different scenarios (including but not limited to campus scenarios, office building scenarios, billiard hall scenarios, large electronics factories, machinery factories, indoor 5A\4A scenic spots, swimming pool scenarios, KTV scenarios, parent-child scenarios, gym scenarios, bathing center scenarios, karting park scenarios, exhibition hall scenarios, bar scenarios, botanical garden scenarios, e-sports hall scenarios, postgraduate entrance examination study room scenarios, large hotel scenarios, board game scenarios, etc.)
[0024] (2) Product information organization Organize the detailed information of the products in the smart vending cabinet, where the product information includes product ID, product brand, product name, product category, product specification, product packaging, product barcode, product price, product display map, inventory, etc. This information helps the algorithm better understand the relationship between products, thereby improving the accuracy of recommendations.
[0025] (3) Define multidimensional array structure In the commodity order data set module of the system, records of consumer purchase orders from different scenarios and channels are collected and displayed. Purchase, evaluation and other behaviors are recorded in the system to form a four-dimensional array of consumers (people) - smart vending equipment (machines) - commodities (goods) - scenarios (fields).
[0026] ①The first dimension: consumer (user) dimension, which records the information of each consumer.
[0027] ②The second dimension: the smart vending device dimension, which records the information of each device and its interaction with consumers.
[0028] ③The third dimension: product dimension, which records the information of each product and its sales in various devices and consumers.
[0029] ④The fourth dimension: scenario dimension, which records consumer behavior, device status and product sales in different scenarios.
[0030] 2. Data preprocessing and construction of interaction matrix (1) Clean the collected data in the four dimensions and remove invalid or redundant records.
[0031] (2) Use mean filling to handle missing values.
[0032] (3) The data is used to construct multiple interaction matrices such as consumer-product, consumer-device, device-product, and consumer-product in different scenarios.
[0033] ① Consumer-product interaction matrix: records consumers’ ratings or purchases of products.
[0034] ② Consumer-device interaction matrix: records consumers’ device usage or preferences.
[0035] ③Device-product interaction matrix: records the sales or recommendation times of products in the device.
[0036] ④ Consumer-product interaction matrix under different scenarios: records consumers’ purchase status or ratings of products under different scenarios.
[0037] S2 system recommendation algorithm module implementation stage 1. Calculate product similarity Select the Jaccard similarity calculation method. Let set A and set B represent the feature sets of two products respectively. The product feature set includes the purchase and sales records, brand, category, browsing history, etc. of the product. The Jaccard similarity calculation formula is: J(A,B)=|A∩B| / |A∪B| Among them, |A∩B| represents the number of elements in the intersection of set A and set B, and |A∪B| represents the number of elements in the union of set A and set B.
[0038] 2. Generate a collection of similar products Based on the similarity of products, a similar product set is generated for each product. This set contains products that are highly similar to the target product.
[0039] 3. Predict ratings and generate recommendation lists (1) Calculate the average rating of similar products for the target user, The steps are as follows: ① According to the feedback mechanism of the S3 module, obtain the user-product rating matrix: Assume the matrix R, where R ij represents the rating of user i on product j; ②According to the collection of similar products generated by the algorithm, the product similarity matrix is obtained: Assume that the matrix S, where S jk Represents the similarity between product j and product k; ③ Target user: Assume that the target user is u. ④ Target product: Assume that the target product is a; ⑤ Extract products similar to the target product a from the product similarity matrix S. And set a threshold θ, only consider products with a similarity greater than θ. ⑥ For each product k similar to the target product a, calculate the score of the target user u for product k; ⑦ Add up the scores of all similar products and divide it by the number of similar products. ⑧ Get the average score of the target user for products similar to the target product.
[0040] (2) Based on the predicted ratings, a recommendation list is generated for the smart vending device merchant, which contains the products with higher ratings. The recommendation list generated by the algorithm provides the smart vending device merchant (operator) with relevant decisions on product selection and supply.
[0041] S3 system platform algorithm optimization and adjustment 1. Update the recommendation list regularly Regularly update the recommendation list based on real-time changes in consumer purchase behavior data and product information. Ensure the timeliness and accuracy of the recommendation results. Provide direction for merchants (operators) of smart vending devices to adjust product selection.
[0042] 2. Introduce a feedback mechanism (1) Set up a feedback button or evaluation system on the smart vending machine or set up a questionnaire and rating of purchased goods to collect user feedback on the recommended product selection results.
[0043] (2) Fine-tune and optimize the recommendation algorithm based on user feedback.
[0044] 3. Algorithm evaluation and improvement (1) Obtain the latest product order records and update the product information data in real time. At the same time, count the types and quantities of products in the product recommendation list generated by the algorithm in the latest product order records, calculate the accuracy, and evaluate the performance of the recommendation algorithm based on the accuracy.
[0045] (2) If the accuracy rate is high, continue the above process to generate a recommended product list; if the accuracy rate is low, increase the number of historical orders obtained, or adjust the product types based on the latest orders, and repeat the above steps to improve the accuracy of the recommendation results and consumer satisfaction, while increasing the merchant's sales and profits.
[0046] S4 system platform optimal supply route module 1. Intelligent vending equipment automatically reports location The freezer location can be obtained through API interfaces such as Amap and Baidu Map. The optimal product supply route can be matched according to the selected equipment platform to save labor costs.
[0047] The above embodiments are only exemplary embodiments of the present invention and are not intended to limit the present invention. Those skilled in the art may make various modifications or equivalent substitutions to the present invention within the essence and protection scope of the present invention, and such modifications or equivalent substitutions shall also be deemed to fall within the protection scope of the present invention.
Claims
1. A human-computer interactive product selection and supply method, characterized in that: The specific steps include: Step 1, obtain the consumer's historical purchase information data, define a multidimensional array structure, and perform data processing on the multidimensional array structure; Step 2: Calculate the similarity between products based on the Item-CF recommendation algorithm, generate a product recommendation list, and introduce new products; Step 3, through the optimal supply route module, generate the optimal route for the supply replenishment point according to the location of the system; Step 4, obtain the latest product order consumption record, update the product information data in real time, and count the types and quantities of the products in the product recommendation list generated in step 2 in the latest product order record, and calculate the accuracy rate; Step 5: After purchasing the product, the purchase experience is rated through the user feedback module to collect consumers’ feedback on the recommended product results; Step 6: Adjust the product recommendation list based on the accuracy obtained in step 4 and the information collected in step 5.
2. A human-computer interactive product selection and supply method according to claim 1, characterized in that: The specific method of defining the multidimensional array structure in step 1 is: setting corresponding scenes for devices placed in different scenes in the product selection and supply system, recording detailed information and purchase evaluation behaviors of the products in the devices, and the product order data set module collects and displays order records of consumers purchasing products from different scenes, forming a four-dimensional array of consumer-device-product-scene.
3. The human-computer interactive product selection and supply method according to claim 1, characterized in that: The data processing in step 1 specifically includes: clearing invalid or redundant records in the four-dimensional array data, processing missing values with mean filling, and constructing the data into a consumer-product interaction matrix, a consumer-device interaction matrix, a device-product interaction matrix, and a consumer-product interaction matrix under the scenario.
4. The human-computer interactive product selection and supply method according to claim 1, characterized in that: The step 2 specifically includes: calculating the similarity of the products by using the Jaccard similarity calculation method, generating a set of similar products, calculating the average rating of the target users for the similar products, and generating a product recommendation list according to the rating.
5. The human-computer interactive product selection and supply method according to claim 4, characterized in that: The specific method for calculating the average score of similar products by the target user is as follows: a user-product score matrix is obtained through a user feedback module, a product similarity matrix is obtained according to a set of similar products, products similar to the target product are extracted from the product similarity matrix, a threshold is set, products with a similarity less than the threshold are eliminated, a target user score is calculated for each product with a similarity greater than the threshold, and the average of the target user scores of all products with a similarity greater than the threshold is calculated to obtain the average score.
6. A human-computer interactive product selection and supply system, characterized in that: A human-computer interactive product selection and supply method based on any one of claims 1 to 5, comprising a product order data set module, a recommendation algorithm module, an optimal supply route module, and a user feedback module, wherein the product order data set module is used to collect and display product purchase orders; the recommendation algorithm module is used to generate a product recommendation list based on historical order data in the product order data set module; the optimal supply route module is used to generate an optimal route for supply replenishment points based on equipment geology; the feedback module is used to collect user feedback on product selection and optimize the recommendation algorithm model based on the feedback.
7. The human-computer interactive product selection and supply system according to claim 6, characterized in that: The Item-CF recommendation algorithm calculates the similarity between products and generates a recommendation list based on the similarity of the products and the historical behavior of consumers. Merchants select products for sale based on the smart list, collect the latest order records of consumers purchasing products, and count the products and quantities purchased from the recommendation list.
8. An application of the human-computer interactive product selection and supply system as described in any one of claims 6-7 in a smart vending cabinet.