| Data mining has become a commonly used tool for different fields. Statistics has played an important role in data mining, from data exploration to analysis. Not only are its theories used as solid foundation, but its methods are integrated into data mining.;In this thesis, statistics and its application in data mining are discussed, such as statistics models of mixture model, regression, Bayesian theorem, along with methods based on them, such as mixture model based clustering, Bayesian theorem classifier, and regression based anomaly detection.;Data mining projects in Saskatchewan Public Health claims are used to demonstrate statistical methods. Furthermore, results from statistical models are compared with results from other methods. It is found that although statistics models have their unique advantages, they are also subject to some limitations. Therefore, combining statistical methods with other methods can yield more robust results. |
