Summary: | 碩士 === 國立中正大學 === 數學系應用數學研究所 === 107 === This paper applies the methods of machine learning in business data from e-commerce, travel service, or insurance companies.Although these companies belong to different industries, they have the same purpose of reducing internal operating costs via shopping data analysis and consumer behavior forecasting. Recently, the speed and quantity of data collection is certainly far faster and larger than the past. Therefore, not only the traditional statistical Regression Analysis, Logistic regression, and Principal Component Analysis, but the machine learning methods developed in recent years, such as Neural Network, Support Vector Machine, Random Forest, XGboost are used in this paper. Through the comparing and discussing between these methods, we build up the forecasting models, search the key factors in the models, and interpret the observed phenomena.
|