With the rapid development of Chinese economy, a great deal of data from long-term management processes has been preserved in modern enterprises. It will be important for enterprises management decision, if the large customers' data are explored. Data mining (DM) technology has become a popular tool turning those data into valuable information and knowledge, and Customer Relationship Management (CRM) has become an important application field of Data Mining.In this thesis, some new statistical methods-Multivariate Additive Regression and Tree (MART), Markov Chain Monte Carlo (MCMC) etc, are put into complex data mining based on characteristics of complex data in service industries. A statistical analysis frame of complex data mining with high creativity is formed, and some mining steps of complex data in practical service industries by using these statistical methods are discussed.Meanwhile, a great contribution in aspects of customers' segmentation and customers' vialues of simulating prediction by using MART & MCMC methods respectively from a typical service industry-hotel customer information and data, is made. Better ;analysis results by using these statistical methods than_classical statistic methods are obtained; it is important to practical contribute and be feasible for CRM in practical enterprises in aspects of conclusion explanations and decision-making suggestions, and a scientific decision-making foundation of enterprises effective marketing tactics is provided.
|