台灣地區觀光旅館之營運績效採礦研究

Data Mining of Operation Performance of Tourist Hotels in Taiwan

在二十一世紀，世界各國普遍將觀光產業視為明星產業，且隨著現代人對生活品質的重視，人們對休閒活動的需求亦跟著增加，其中觀光旅館可供民眾住宿、餐飲、會議、購物等功能，為了解觀光旅館經營績效背後隱藏的因素，針對影響觀光旅館營運績效的重要因素評估。本研究使用2003年到2012年觀光旅館之營運統計，利用決策樹模型和多元迴歸模型等資料採礦方法，探索影響觀光旅館營運之重要因素，挖掘營運管理之決策規則，並建立總營業收入之最佳預測模型。研究結果發現，考慮星等、餐飲收入、住用率和員工人數之線性迴歸模型，解釋能力（R_Adj^2）可高達92%（模型1）；若是簡化模型只以星等和員工人數估計觀光旅館之總營業收入，此線性迴歸模型之解釋能力（R_Adj^2）可達84%（模型2）；因旅館營運績效存在影響性資料，若是刪除模型2之影響點後，調整後判定係數（R_Adj^2）將可達到89%（模型3），赤池資訊量（AIC）為-1373.42，驗證之平均平方誤差為0.08，預測誤差（PRESS）為27.18；其中尤以五星級觀光旅館之平均營運績效為最佳。旅館業為勞力密集產業，需要投入大量人力，消費者對於員工之服務品質相對要求較高，若能提升員工之服務品質，應有助於經營績效之提升。

With the modern emphasis on quality of life, the demand for leisure activities also followed increases. Tourism industries are usually considered significant industries around the world. Tourist hotels could provide the multiple functions such as public accommodation, dining, meetings, shopping and so on. This study would like to understand the key factors affecting the operational performance of tourist hotels in Taiwan. The operational database of hotel industry from 2003 to 2012 is cited from the Tourism Bureau, M.O.T.C. The results indicate the hotel’s magnitude and total number of employees would significantly affect the total operating revenue for the tourist hotels (R_Adj^2=84%). Five-star tourist hotels have the best operating performance. If adding occupancy rate and catering revenue into the prediction model, its ability of explaining total revenue as much as 92%. The hotel industry is labor-intensive industries. It needs to invest a lot of manpower and to provide the high-quality service of employees for their customers. As the management are devoted to improve the service quality of employees, it is helpful for the promotion of operational performance of hotel business.