Analysis On Data Mining Model Based On The Consuming Behaviors Of Customers In Value-Added Business

With the gradual increase in the proportion of low-end users, the market competition is becoming increasingly fierce in the mobile communication industry. The revenue per user ARPU (Average Revenue Per User) of the traditional businesses such as voice business continued to decline, new businesses are necessary for the telecom operators to maintain or even increase ARPU. Mobile Data Service provides the opportunity for the operators. Mobile data traffic provides this chance for them .However, how to use mobile data traffic to increase ARPU for the customers is becoming a burning question to the telecommunications value-added service. On the one hand, the irregular operation of SP (service provider) had a strong impact on the image of good faith of telecommunications operators. The problem is partially improved in the rectification on SP and CP (content provider) in August last year. On the other hand, business telecommunications value-added services use outside the normal channels to call users and PUSH messages to recommend new businesses, response rate of the users is relatively low . And they also brought a large load to the network. It has caused the increase on invisible operating costs. More importantly, some very good businesses were recommended to the inappropriate people, which caused the complaints from some users. How to settle such problems is the main topic of this paper.The CRISP-DM (Cross-Industry Standard Process for Data Mining) is used in this paper, which is one of generally recognized influential methodologies. CRISP-DM stressed that DM is not the organization or presentation of data and also not data analysis and statistical modeling but a complete process from understanding of business needs, seeking solution to meet the test of practice. The entire CRISP-DM mining process is divided into six phases :①Business Understanding) ;②understanding the data (Data Understanding) ;③data preparation;④ Modeling ;⑤Evaluation and⑥deployment. A knowledge database of customers' consuming behavior is constructed by applying these technologies to analyze the consuming behavior of the customers. The knowledge database of customers' consuming behavior is used to infer the preferences of the customers by using a series of algorithms and analysis on the basis of the consuming records.The data mining software Clementine of SPSS is used in this paper by applying related analysis to analyze a large amount of historical data from point-to-point messaging, MONTERNET and WAP businesses to find the related relationship between the choices of the consumers on mobile data services. It can find the consuming features of consumers and then choose relevant consumers according to the business profiling. Precision can be improved and marketing costs can be reduced by outside call and PUSH message. The respond rate of the consumers also can be increased. The object of improving the ARPU of the users can be achieved at last.The methods used in this paper to improve the efficiency of the spread of mobile value-added service that has made initial success. Network marketing has spread to many similar experiences, which is the principal value of this paper.