A Study On Potential Customers Finding Model Based On The Behavior Characteristics

While the continuous developing of the Electronic Commerce, the number of products isincreasing rapidly and the message is expanding on the Electronic Commerce net. According to that,there are two difficult problems appeared during the developing of the Electronic Commerce. Onthe one hand, due to too many choices, customer feels vulnerable to scan one by one. On the otherhand, it has been the core of management in dealing with the relation between customer andcompany in Electronic Commerce companies to satisfy the special requirements of customer, toraise the loyalty of customer and to discover potential customer. Nowadays, there still have defectsin the discovering potential customer on various net. It is the main problem to solve that is discoverpotential customer accurately.This paper summarizes the present situation of overseas and native discovering customer andconsequently arrives at the clues——the research of model of discovering potential customer basedon the behavioral characteristic. The relation between customer’s behavioral characteristic andkinds of customer is actually a classification problem. Rough set, which does not require any formermessage not related to the objective problem before analyzing, is a new developing classificationmethod. It can discover hidden knowledge and potential discipline. This paper does advancedanalysis on the data of customer’s behavior, including data purification, customer recognition, visitsstatistics, parchaseing times statistics, customers data matching, attribute recognition of whether tobuy after browsing, redundancy data deleted, research object selection. In order to increase theaccuracy of discovering potential customer and reduce the disturbance of other kinds of customer,this paper choose the customer who buy products once or just browse but not buy. Additionally, thepaper researches the excavation model of customer’s behavioral characteristic based on tough set.This part includes three steps, such as customer behavioral data scattering, customer behavioral datasimplification and customer behavioral characteristic collecting. In the step of customer behavioraldata scattering, this essay introduces scattering algorithm, which is based on message entropy, tomake the compatibility unchanged after scattering. In the step of customer behavioral datasimplification, this paper introduces characteristic simplification algorithm based on messageentropy on account of thinking of system’s compatibility. In the step of customer behavioralcharacteristic collecting, this essay uses the discipline generation algorithm based on confidencedegree to make the generated discipline be more real and crucial. After the depthly analysis for thelarge-sample of the customer behavior data and the otherness among the different types ofcustomers, this paper improving the extraction algorithm of behavior rules. Following that, this paper establishes the potential customer discovering model based on behavioral characteristic. Thefirst step is to design the pretreatment process of the data. The second step is to establish acharacteristic database based on the customer’s behavioral characteristic which is got according tothe behavioral characteristic collecting algorithm based on tough set. The third step is to advancedprocess and scattering to the customer’s behavioral data which needs to forecast. The last step ismatching the access behavioral data after advanced process and the data in behavioral characteristicdatabase to discover whether the one is potential customer.The analyze of experimental result reveals that the model of discovering potential customer,which is based on behavioral characteristic, can significantly increase the accuracy of discoveringpotential customer. Finally, we do summery on the research in this paper, analyze the drawback ofthe model and make the prospect of further research.