Design And Implementation Of Second-hand Commodity Trading Platform Based On Intelligent Recommendation And WebGIS

With the explosive growth of the scale of the second-hand commodity market,the traditional second-hand commodity trading platform cannot meet the personalized needs of current users,and it is difficult for users to quickly find the commodities they are interested in and convenient for offline transactions in the massive commodity market.Therefore,it is of great practical significance and theoretical research value to provide users with personalized commodity recommendation services.Based on the optimized collaborative filter intelligent recommendation algorithm and WebGIS service,this project designs a C2 C mode second-hand commodity trading platform,realizes the functions of intelligent commodity recommendation and shortest-term trading route planning,and provides users with more personalized recommendation services.The system uses Vue,Express,Node.js and other development technologies and frameworks to develop intelligent recommendation functions and management functions for users,commodities,transactions,and payments;the system implements dynamic map loading based on Open Layers technology,and applies Dijkstra’s shortest path planning algorithm to recommend the shortest-term transaction route for users;the system uses Java Script technology to realize automatic identification of book ISBN barcodes and three-party payment functions.In order to better solve the problem of accurate recommendation of goods,this project proposes a collaborative filtering intelligent recommendation algorithm that integrates the attribute factors of the project,and applies it to the intelligent recommendation service of goods.First,collect user rating behaviors and extract commodity attribute keywords.The user’s browsing,purchasing and other behaviors are stored in the form of ratings in the usercommodity rating table,and the commodity attribute keywords are extracted through word segmentation and TF-IDF methods.Second,a weighted similarity matrix is calculated based on the commodity attribute factors.Second,a weighted similarity matrix is calculated based on the commodity attribute factors.The commodity similarity matrix based on the usercommodity score is calculated,the commodity similarity matrix based on the commodity attribute is calculated,the commodity similarity matrix based on the user-commodity score and the commodity similarity matrix based on the commodity attribute are linearly combined according to a certain weight to obtain the commodity similarity matrix for recommendation.Finally,build a recommended commodity set based on the predicted interest commodity score.Based on the user’s past interest commodities and the weighted commodity similarity matrix based on the attribute factors of the commodity,the score set of the predicted interest commodity is calculated,and the recommended commodity set is constructed by arranging the score set in reverse order.Through intelligent recommendation algorithm and WebGIS technology,this system provides users with more humanized intelligent commodity recommendation services and offline shortest transaction route recommendation services,which has positive significance for improving the user experience of the system and increasing the stickiness of platform users.