Application Of Bayes Classifier Based On Clementine

In the era of big data, how to deal with the numerous data has become a hot issue in today’s society. Bayesian network, proposed at the end of the last century as a data processing tools, combines the advantages of the graph theory and probability theory, and has shown its superiority in the field of uncertain information, for which is being used for the data mining and analysis by more and more experts.This thesis aims at learning and applying the Bayesian network. It introduces the theory of three types of Bayesian network classifier from the structure learning and parameter learning. Then, the thesis describes and analyses the current data processing methods and issues of mobile Internet industry. There are a lot of waste of data in the current mobile Internet industry and many important data are not fully utilized. They have no awareness of using data to make decision. At the end, it constructs two Bayesian network models to classify the different kinds of users. Both of models have a reasonable explanation. It makes the mobile Internet companies capable of predicting possibility of loss and subscribers’ future paying ability, and organizing promotion and activities corresponding characteristics of different users.From the results of the case analysis in this thesis, we can see the applicability of the Bayesian network classifier and the importance of Clementine in the data mining. In the future work of the Internet, we can utilize more data mining methods, analyze and mine data with software from a new perspective, and find important information hidden behind the data.