The Design And Implementation Of A Store Recommendation System

With the rapid development of Internet and 4G networks,network traffic flows from the PC to the mobile phone side.The company have accumulated a large amount of data.In order to achieving the win-win strategy,how to use the accumulation of consumer behavior information and geographic information to provide users with personalized recommendation,improve customer's satisfaction,at the same time make the company's profit growing,is the problem which needs to be settled.Our recommendation system consists of two parts: the recommendation algorithm engine and the background visualization module.The recommendation algorithm engine is the core module of the system.Recommendation engine by analysis the user's consumer behavior and combined with the user's geographical information to recommend products for users.For different user groups and behaviors,the engine generates a recommendation list for the user by launching a different algorithm engine and strategy.In order to improve the performance of recommend engines,we use the Wilson interval algorithm to process the shop scoring matrix,and we introduce CSR to compress and store the scoring matrix.At the same time,in order to improve the efficiency of similarity calculation,we introduce the ball tree.That improve the efficiency of computing retrieval.In the system implementation,we make a preliminary analysis of the user's geographic information,in the recommendation engine to introduce geographic information,so that the final recommendation more personalized.In order to help our facilitate the better understanding of the data and the operation status of the website by the algorithm engineer,the operation,the product manager,etc.Our system through the establishment data analysis engine,carries on the user behavior data,the shop Consumption,regional traffic and so do a statistical analysis,in order to speed up the efficiency of data to show,we set the timer to trigger the script,regularly update the database.