Analysis And Research Of Online Shopping Users Based On Data Mining Technology

Data mining technology is now widely used in business analysis and processing.With the continuous progress and development of social science and technology,people in modern society generally have a higher pursuit of material life,and their material needs are more diversified.On the contrary,the market supply often fails to meet the needs of the majority of users.Therefore,the market often appears the imbalance between supply and demand,most products will have a short supply or oversupply phenomenon.Therefore,for enterprises,analyzing and understanding the behavior characteristics of each user and user group,including users’ purchase preferences and purchase habits,can provide users with better services and diversified products.This can not only help enterprises gain benefits but also improve user satisfaction,but also break the imbalance between supply and demand.The paper is the analysis and research of the purchasing behavior of online shopping users based on data mining technology.The paper first describes the research background and research status of online shopping user behavior in detail,and understands the relevant theoretical basis of data mining technology and user purchasing behavior in detail.The data set of online shopping user behavior provided by Ali Tianchi big data platform used in the paper.The paper divides the analysis of user behavior of online shopping into two levels: one is to analyze the characteristics of user purchase behavior;the other is to recommend the product according to the purchasing behavior of the first step.Accordingly,first,a new model as the BCFPT classification model.RFM has only have three value indicators to classify users,which contains limited information,insufficient interpretation of business problems,and RFM classification will be disturbed by outliers.In this case,calculating the average of each index will lead to inaccurate classification boundaries in other groups,add new indicators to classify the users of the platform,design and propose a new model as BCFPT classification model,so as to improve the accuracy of classification,and make user value analysis according to the displayed classification results and platform business characteristics.Second,the speed of FP-growth algorithm is faster than that of Apriori algorithm,but when the data amount is large,FP-growth algorithm finds sets frequently,and its excessive child nodes will reduce the algorithm efficiency.Therefore,the paper proposes the improved D＿FP-growth algorithm to delete meaningless data items and increase the algorithm efficiency,so as to achieve the purpose of recommending suitable products to users.