Research On Personalized Recommendation Algorithm And Its Application In Intelligent Retail System

With the rapid development of the economy,the real estate industry has also shown a trend of development,resulting in a substantial increase in rents in prosperous areas,and domestic retail sales have been greatly affected.The aging of the social population has led to an increase in labor costs.In recent years,the convenient,fast and low-cost vending machine industry has developed rapidly.At the same time,people's pace of life is accelerating,and users' shopping time is becoming more and more fragmented.As the consumers' notions of individuality begin to be widely pursued,the traditional vending machine industry is also facing a lot of challenges.In order to meet the personalized needs of consumers,only by making personalized sales plans can the vending machine industry develop for a long time.Therefore,how to recommend products that may be of interest to consumers in a variety of commodities and save consumers' time is an urgent problem to be solved.This paper discusses the basic concepts,classification and framework of the recommendation system,compares the common recommendation algorithms,and focuses on the user-based collaborative filtering algorithm based on analysis of domestic and foreign research status.Aiming at the possibility of over-predicting similarity,a correction factor is introduced,and a user-based collaborative filtering algorithm for optimizing Pearson similarity calculation is proposed.This paper analyzes the necessity of introducing item-based collaborative filtering algorithm,introduces its basic concept,and proposes an optimized hybrid collaborative filtering algorithm.The design and implementation of the experimental scheme were carried out in combination with the specific data set,and the feasibility and accuracy of the improved recommendation algorithm were verified.Based on the above improved algorithm,combined with the needs of the vending machine,a recommendation system was designed to realize the personality recommendation of users and the recommendation management function of the administrator.Finally,the system interface and operation flow are given,which proves that the personality recommendation system has high practical value.