Data Mining in Churn Analysis for Hotel Customers

• Main Authors: TSAI, YU-RUI, 蔡育叡
• Other Authors: LEE, YUE-SHI
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/un4vn5

碩士 === 銘傳大學 === 資訊工程學系碩士班 === 107 === With the development of the online world and the increasing scale of online users, the online booking hotel market has great potential to be discovered. Due to the increasing size of the booking market, it has gradually attracted the attention and use of local people. According to the data, the market share of Ctrip.com currently accounts for more than 50% of the market in China's online hotel bookings, which is quite large. It is of great significance to study Ctrip.com, a leading online travel booking service company in China, to identify the key factors of its user loss and prevent the business crisis caused by customer loss. Based on the theory of customer churn, this paper first preprocesses the data and then uses the existing features to create new features. Then it uses the XGBoost algorithm in the data mining method, the random forest algorithm, and the neural network algorithm, constructing the model of customer churn based on Ctrip.com data, and evaluating the performance of these models. Secondly, according to the optimal model constructed, the key factors leading to the loss of Ctrip customers are analyzed. Finally, based on the results of the analysis, suggestions for reducing the loss of Ctrip customers are proposed.