國際觀光旅館之住宿人數預測模型研究

The Study of Forecasting Model on Tourist Arrivals in the International Tourist Hotels

徐振倫

指導教授 ： 方近義

國立臺灣師範大學/運動與休閒學院/餐旅管理與教育研究所/碩士(2011年)

摘要

觀光產業影響所及餐飲、旅館、運輸、民生消費等產業，可帶動整體相關經濟發展，為一國家之軟實力，其重要性不可言喻。2009年開放兩岸直航、2010年松山機場直飛上海虹橋機場及日本羽田機場，國際旅客倍增，許多企業都看好台灣市場，紛紛興建大型連鎖旅館。
本研究針對六家星級國際觀光旅館，以時間序列ARIMA與向量自我迴歸模型VAR方法，除了以歷年住宿人數為預測基礎，突破單一變數的影響，加入影響旅館業住宿人數的三項總體經濟變數(國內生產毛額、消費者物價指數及匯率)與旅館特性變數，實證探討比較不同方法對星級國際觀光旅館之住宿人數預測模型之建立。資料來源為交通部觀光局公佈之國際觀光旅館住宿人次統計資料及行政院主計處「中華民國統計資訊網」，資料選取為從2000年一月至2010年十二月，共132筆月歷史觀察值。
透過上述方法，預期成果能針對個別之星級國際觀光旅館建立住宿人數預測模型，並加入總體經濟因素分析及旅館特性變數，提高模型預測能力，以MAPE為模型選擇標準，所得結果，俾作星級國際觀光旅館業者及相關單位之施政參考依據。

關鍵字

差分自我迴歸移動平均 ； 向量自我迴歸模型 ； 國際觀光旅館 ； 預測

並列摘要

Taiwan’s tourism industry is getting more prosperous and booming after the policy deregulation for Chinese tourists to Taiwan since July of 2008. Tourist arrivals and expenditures from Chinese tourists had brought more business opportunities to Taiwan. Besides, plenty of international chain hotels are under construction. That means world investors have faith in Taiwan’s hotel industry. Hence, an accurate tourism forecast is particularly crucial not only to governments and practitioners but also to investors’ resource allocation and decision making. Tourism demand forecast has been widely explored in recent years. The main objective of this study is therefore to obtain more accurate forecasts of Taiwan specific international hotel total arrivals by comparing ARIMA and VAR model, which are rarely employed in hotel industry. In this study, monthly data covered the periods from 2000/01 to 2010/12. The data period from 2000/01 to 2010/03 was used to build the forecasting model, while the remaining data was used to evaluate it. The data is collected form The Monthly Report on Tourism published by the Tourism Bureau of Taiwan. Explanatory variables included gross domestic product, consumer price index, exchange rate, and hotel characteristic variables. In order to evaluate of the proposed modeling performance, we use the mean absolute percentage error (MAPE) as evaluation. Results show that MAPE is lower than 20%, indicated model has good forecasting ability. Better performed forecasting model has more favorable benefits toward the resource scheduling, capacity planning in the management team of the international tourist hotel.

並列關鍵字

ARIMA ； VAR ； International tourist hotel ； Forecasting

參考文獻

陳宛君、鍾宜陵，(2009)，國人於台灣及國外地區觀光旅遊需求之研究，戶外遊憩研究，22(2)，1-26。

Armstrong, J. S. & Collopy, F (1992). Error Measures for Generalizing About Forecasting Methods: Empirical Comparisons. International Journal of Forecasting, 8, 69-80.

Choi, I. (1992). Effects of data aggregation on the power of tests for a unit root. Economics Letter, 40, 397-401.

Cranage, D. (2003). Practical time series forecasting for the hospitality manager. International Journal of Contemporary Hospitality Management, 2, 86-93.

De Mello, M. M. & Fortuna, N. (2005). Testing alternative dynamic systems for modeling tourism demand. Tourism Economics, 11(4), 517-537