Applied Research, The Cart Algorithm In The Identification Of Potential Customers Of The Telecommunications Industry

Our society is full of information, lots of data occur every day, so it is a critical question that how can we get useful information from these data and then create more value for us. Data mining bases on researching data, then we can get useful information for the further deciding. These years, more and more companies focus on the data mining, so it is also a researching direction for the data house and information deciding.Data mining has been used in many areas, the techniques include Statistics, Machine Study and Information Science. Now, it is also used in Neural Net, Rough Set and other science.This paper will elaborate the basic theories of data mining after introducing research background, meaning, domestic and international development of this topic, studying different classifications of data mining. Bases on this foundation, we will analyze decision tree theories and basic arithmetic which belongs to machine study category, discuss different branch standards and shear methods.Because there are many decision tree arithmetic, this text will emphasize on CART arithmetic, this arithmetic is a method of non-parameter statistics and mainly being used for classification research, it can deal with both of continuous and classification variables, after the analysis of the whole process and advantage of this arithmetic, this article will do a evidence research about the latent customer in telecommunication industry, bases on customer's one month call data of one telecommunication company, we will use CART arithmetic to build a mode of latent customer identification, after analyzing the result of the mode, we can confirm the characters of the latent customers, if we can carry on a proper marketing work to these customers, ether the company or the customer will both benefit a lot from this.