Research On Agricultural Products Trading Recommendation System Based On Optimized Particle Swarm Optimization Algorithm

Today,with the development of artificial intelligence driven by big data,the redundancy of big data increases the cost of obtaining key information.And the recommendation system has become one of the effective ways for users to obtain key information.The platform for fresh agricultural product solves the problem that urban residents demand fresh fruits and vegetables in their daily life.However,with the continuous development of the platform for fresh fruits and vegetables,it has been more difficult for users to find the ideal one from many platforms.Therefore,this paper builds an e-commerce recommendation system of agricultural product based on the community user environment,and improves the recommendation accuracy of agricultural products merchants by mining user behavior and its' relationship in the community.In this thesis,by undertaking research and making improvements on particle swarm optimization algorithm,the design and implementation of the system are completed through the micro-service technology framework and recommendation algorithm.The research mainly focuses on the following aspects:(1)Starting from the e-commerce function of agricultural and sideline products and the user life scene,this paper analyzes the trading patterns of e-commerce platforms for many agricultural and sideline products,and considers the problems in traditional recommendation algorithms by collaborative filtering such as cold start,data sparse,etc.,and builds a trading recommendation model of the agricultural product based on users' location in the community and their purchase preferences.The model considers the distance and purchase preference similarities of the user,and the star rating similarity of merchants,so as to form a combined recommendation for obtaining the nearest neighbor set.And the PSO algorithm were taken with the nearest neighbor setting as the initial particle group.In the PSO algorithm,particles quickly search for global extrema by constantly sharing individual extrema of their current position.The global extremum is transformed into the initial population by genetic algorithm,and the global convergence ability of the genetic algorithm is exploited too.The crossover and mutation ability of the genetic algorithm is used to make up for the shortcomings of the particle swarm algorithm which is tend to fall into the local optimum,and all of that has proved the advantages of the model based on the particle swarm group optimizing algorithm.By comparing the two test indicators of MAE and Precision with that of the traditional recommendation model,the accuracy of the improved recommending model is verified and the expected goal of this paper is achieved.(2)The e-commerce recommendation system of agricultural product adopts the B/S three-tier architecture and the SpringBoot+Dubbo micro-service development framework as the overall technology architecture of the system.And the front-end technologies of JavaScript+jQuery,MySQL database and Nginx+Tomcat server have jointly contribute to the design of the system,and the recommendation function of the agricultural product merchants were realized too.This paper proposes an recommendation model of agricultural product based on optimized particle swarm algorithm,and the model realizes the personalized function of recommending process from agricultural and sideline products merchants to their users.In order to verify the accuracy of the model,an e-commerce recommendation system of agricultural product was set up.And the e-commerce recommendation structure that meets the market demand was designed,so as to fully realize the front-end service and the background management functions.