Design And Implementation Of Personalize Intelligent Recommendation System Based On Big Data In Smart Trading Area

In this era of big data,the development of e-commerce continues to impact the traditional retail business.People's consumption habits have gradually changed from the traditional consumption of shopping circle to online shopping.In order to cope with this change,we should vigorously develop the economic zone based on the intelligent trading area.We need an application which can be able to integrate the traditional trade circle resource and service to enhance the service level of the trading area,giving users better service and consumer experience.In this paper,the research and implementation of personalized intelligent recommendation system based on the big data of intelligent trading area.The main aspects of personalized intelligent recommendation system in this paper are as follows,(1)Describes the background and objectives of the research,and outlines the structure of the paper.(2)The main technologies used in the recommendation system are analyzed and introduced.For example,massive data storage used by the HDFS file system,Hive data warehouse,big data distributed computing framework MapReduce,the machine learning algorithm library Mahout.(3)This paper describes the application scenarios of the intelligent business circle,analyzes the application model,user model and data model of the recommendation system.The influence of these three models on the design of the recommendation system is also discussed.(4)According to the application scenarios in the analysis of the intelligent trading area,personalized recommendation system requirements and architecture,and the recommendation system calculation function consists of data collection,offline calculation and online calculation of three parts.The data collection part is responsible for access to businesses or commodity data through the ERP background;the offline part is responsible for the collected data processing and cleaning,similarity calculation and modeling,constructs a semantic chain;online recommendation based on offline part is responsible for processing data to calculate the final recommendation results(5)Describes how to use Hadoop and Mahout to achieve personalized recommendation system.The test results show that the system can be applied to many scenes in the trading area,providing an accurate personalized recommendation services.The recommendation technology used in this paper has some reference value and prospect.