Research On E-commerce Big Data Mining And Application Based On Association Rules

With the rapid development of Internet,big data,artificial intelligence and other technologies,online shopping has become the main consumption behavior of people,and consumers have higher and higher requirements on the quality of goods and delivery time of e-commerce.At the same time,in order to improve the market competitiveness,in addition to improving the quality of service,e-commerce should also cater to consumers’ preferences and shopping behavior,and then implement accurate marketing and fast delivery.Therefore,how to grasp consumers’ behaviors and preferences and make accurate product recommendations;How to optimize the selection operation and shorten the delivery time of shopping will become the key and difficult point for e-commerce to improve operation efficiency,increase marketing profits and strengthen competitive advantages.It is also a research hotspot of scholars.Based on the analysis of e-commerce demand,this paper takes e-commerce shopping big data as the research object and uses big data mining,association rule algorithm and other methods to analyze and study consumer consumption behavior and expand the application field of association rules.Aiming at three important problems in the e-commerce industry,namely,precise marketing,optimization of warehouse space allocation and optimization of order allocation,this paper proposes an optimization scheme of mining similar customer association rules by combining RFM customer value analysis model,K-means clustering analysis and FP-growth algorithm.The design scheme is deployed in Spark platform for experiments.Verify the feasibility of the scheme and the advantages of big data mining in Spark computing framework.In the current electricity for distribution of goods and the order partial solutions based on the study of the relevant application study,first of all for customer shopping information is extracted,then the customer information on the correlation analysis,the correlation coefficient between the goods and commodities portfolio is closely related to purchase frequency,finally will be the results of the analysis and correlation is used in the design of goods distribution solutions and sorting order.The results show that the research method and design scheme meet the design requirements and have certain reference value for e-commerce enterprises.The main results of this paper are as follows:(1)After mining association rules,it is found that the product recommendation problem that only uses association rules ignores customer characteristics,and the designed product recommendation system is of general accuracy.In this paper,clustering algorithm and RFM model are used to mine similar consumer groups,FP-growth algorithm is used to mine association rules in the same consumer group,and Spark platform is used to accelerate the mining process.Experiments show that the association rules after clustering analysis are more targeted to customer recommendation,and Spark platform has strong computing power and improved computing efficiency.(2)The location optimization model is established through the optimization of commodity correlation degree and the purchase frequency of commodities.With the goal of minimizing the total sorting distance,the goods with higher purchase frequency and higher mutual correlation are stored in close distance to improve the selection efficiency.The basis of the sorting order optimization model is the use of the location layout optimization algorithm,commodity correlation degree is related to commodity location allocation,so there are the same or high correlation degree of goods in the order,the associated order is allocated to the same batch of sorting operation,can effectively improve the efficiency of sorting operation.