The Analysis Of Online Group Buying Behavior Pattern Based On Data Mining

Although Online group buying as a fast developing business mode,its profits increase slowly,so determine the factors that influence online group buying,develop the consumers' behavior pattern,and predict the successive behavior,which can make a big contribution to provide satisfying goods and services to consumers and make suggestions for sellers' saling,operation and management.This paper make the prediction model for online group buying to predict the different type of consumers' buying behavior.Firstly,on the basis of the literature of online group buying and the circumstances of real website,this paper determines the influence factors from the angles of customer,seller and society,and considers time,the characteristics of website and the influence of friends' shopping experiences that was different from others' research findings.Secondly,this paper collects the questionnaires and classifies the online group buying consumers into three classes to analyse the different categories' buying haviour pattern based on RFM model,which is recency,frequency and monetary.The three consumers' classes are important keep consumers,general important consumers and worthless consumers.Moreover,applying the PAM algorithm to solve the K-Mediods clustering.Thirdly,applying the improved C4.5 algorithm to make the behavior prediction model for online group buying consumers.Primary,applying the attribute reduction algorithm of decision tree based on dependency degree to reduce attributes,which use the attribute importance as the heuristic information.After that,using the reduced attributes as the input attributes and the consumers' categories as output attribute to make the decision tree model.In order to simplify the model and improve the accuracy of classification,the paper use the cost matrix,EBP pruning and boosting technology to optimize the decision tree.Then applying the improved model to predict consumers' behaviors and evaluating the accuracy and complexity of the model by the the number of nodes and classification errors.At last,explain the classification rules and find out that the price is the most important factor that can influence online group buying.Furthermore,in the different rage of price,the focus factors are different for different type of consumers,which can better guide the sellers' marketing and management.Finally,applying apriori algorithm to find out the potential and valuable association rules for high value consumers in the shopping goods,the purchased website and the focus characteristics of products,which can provide valuable references for sellers to cultivate the loyalty and dependence of high value consumers.