The Research Of Consumer Behavior Base On Data Mining

With the progress and development of social, the enterprise tends to theglobalization and diversification, the competition among enterprises is very strong,thus the enterprises have to change the focus on the core of custom from the center ofproduct, because, the customs have different selections in consuming process, thequantity of consuming product directly affect on the existence and development, sothe custom as a intangible assets have concerned by each enterprise and become intothe primary factor of existence and development of enterprise. Thus, analyzing theconsumer is the primary object considered for current the management of enterprise,which not only is beneficial to the long development of enterprise and improve thecore competition of enterprise, but also can avoids the enterprise running risk in thecondition of unstable market.According to the current situation, the most research achievements aboutconsumer analysis mainly focus on the diving consumer and the values of consumer,then,can not been provided with the decision-making aid for the enterprise; On theother hand, the consuming about custom have variety of ways, the consumingdatabase of custom excessively increases, the features of consuming have beenchanged, so, in order to resolve the probable, the research in this paper is as follows:Firstly，Analyzing the purchasing behavior process from the point of psychologySecondly，Established forecasting data miming model of consuming behavior;Thirdly，Proposed the feature weigh method based on neural network about howexact the data basis feature about consuming;Fourthly，Proposed the association rule classifying algorithm about how toclassify different type consumer;Fifthly，Established Na ve Bayesian method based on fuzzy attribute denote forthe consuming behavior to forecast;Sixthly，Analyzed the proposed the model and methods in this paper in detail, thecorresponding results were be obtained. The study focuses on consumer behavior by the methods of associationclassification and Bayesian. We can easily see from research content that the logicalrelation of this paper was that the proposed consumer mining model was the wholeprocess analyzing consumer, in which, firstly, pretreatment of consumer data;secondly, exacted the important feature of consumer in order to improve analysisefficiency; thirdly, association classification of consumer, the purpose ofclassification was to reason custom consuming based on Bayesian, then theconsuming situation of each kind of custom could be obtained; Finally, analyzed theBayesian method of custom consuming.