隨著旅遊業興盛,帶動了旅遊需求研究的盛行。其重要性在於運用資訊衡量、預測旅遊業發展,以幫助各相關部門因應可能產生的挑戰,因此吸引學界及各國政府重視。澳門雖僅是亞洲一個彈丸之地,但旅遊業一直是澳門首要產業,對澳門經濟影響甚鉅,惟相關旅遊需求預測文獻不多,本研究即期望能作出精準預測以對澳門旅遊需求研究有所貢獻。 本論文採入境旅客人次作為預測目標,樣本期間為澳門統計暨普查局公布之1991年1月至2007年12月之月資料,使用時間序列分析,依據樣本特性選擇二次曲線模型作旅遊需求預測。 實證分別以最小平方法及Cochrane-Orcutt兩步驟遞迴估計法估計二次曲線模型。進行二次曲線模型的樣本內外預測後,另外將模型與簡算法(Naïve Model)之預測結果比較,以預測誤差指標-平均絕對誤差百分比(Mean Absolute Percentage Error, MAPE)、均方根誤差(Root Mean Square Error, RMSE)及Diebold-Mariano檢定來評價、並比較各模型預測能力。最後將預測結果與過往文獻比較,結論發現在澳門旅遊需求預測上,二次曲線模型有較準確之表現。
The thriving development of tourism stimulates the need for tourism research. Tourism research reveals the significance of utilizing information to measure and forecast the development of tourism, which enables related sectors to deal with possible challenges. As a result, the academic circle and the governments all over the world value the study in tourism. As a tiny area in Asia, Macau relies on tourism heavily, and tourism industry also plays an important role in the Macau’s economy. However, the scientific literatures which predict the demands for Macau’s tourism remain scarce. Consequently, this research is expected to offer an accurate prediction and make a contribution to the research in Macau’s tourism. This study targets tourist arrival and uses monthly time series data from January, 1991 to December, 2007 compiled by Macau Statistics and Census Bureau. In light of the characteristic of the sample data, quadratic model is adopted to forecast tourism demand, using Ordinary Least Squares (OLS) and Cochrane-Orcutt iterative procedure to estimate model. Afterwards, Mean Absolute Percentage Error (MAPE), Root Mean Square Error (RMSE) as criteria, and Diebold-Mariano test are adopted for evaluating the accuracy of the quadratic model. Finally, the Naïve method and the forecast result of past literature are compared with the model adopted in this study. In conclusion, the accuracy of quadratic model in Macau’s tourism demand is significantly better.