The Research Of E-commerce Potential Customers Mining Based On Customer Behavior Analysis

With the rapid development of information technology and the continuousdevelopment of e-commerce market, more and more consumers prefer commodities totrade platform in the network, so the online shopping has become an important way ofshopping in society. And then e-commerce has entered the era of big data, whichgenerated a huge amount of customer data every day, at the same time with a strongdesire of the business knowledge for massive data extracted, how to extract usefulinformation from massive data and help companies develop appropriate decisions, itis particularly necessary and urgent. Currently, as the increasingly competitive ofe-commerce market, which continue to develop more new customers and grasp thedynamic needs of customers, mining of the potential customer groups from the manysuffers that provide personalized service and convert potential customers to become areality customers, these companies will get more benefits and advantage of marketcompetitive. The purpose of potential customers that is to provide accurate referenceand make the appropriate decisions for the website develop appropriate servicestrategy, and currently mining potential customers of many e-commerce sites havesome effect, but it isn’t perfect overall. How to efficient, accurate, and timely to minethe potential customers which the e-commerce companies need to solve the problem.This paper focus on analysis of customer behavior information on e-commercesite, and combines with customers browsing behavior and customer buying behaviordata for a source of data mining potential customer; base on this, using Data Miningof Rough Set and Decision Tree combines with hybrid algorithm to mine on thepotential customers of e-commerce, which to achieve be efficient, accurate and timelygoal to dig out the potential customers. The main contents are as follows:Firstly, this thesis dose research on studies of customer behavior analysis,mining potential customers and so on, which analyzed and summarize by reading alarge number of domestic and international literatures, to accurate knowledge ofdomestic and international research profile and progress.Secondly, with the systematic study of customer behavior on e-commerce website,this mainly on both of customers browsing behavior and purchase behavior onconducted in-depth discussion analysis and understanding, then summarizes the characteristics of customer behavior data. Based on this, the thesis prose to customerson a commodity customers first purchase or not to purchase after browsing as miningobjects, and choose to use Web logs customer browsing behavior and purchasingbehavior for data mining.Then, analysis of Rough set and decision tree about relevant theoretical ofknowledge and the whole process of mining potential customers, and for set ofattributes for the potential customer behavior under large-scale cases, it propose adependence improved methods of attribute reduction. Then based on the behavior ofattribute reduction, it also proposed a new test to distinguish between the value ofmultivariate methods of improving tree structures,that which building the miningmodels of e-commerce potential customers.Lastly, this paper uses an e-commerce website data and preliminary researchdata to propose with analysis and simulation evidence. The experimental results showthat this method can construct an ideal of decision tree model, and extraction thebehavior rules is more easier to understand and interpret, which can accurately dig outof the potential customers.