根據世界觀光組織（World Tourism Organization, WTO）分析 報告指出，在全球各國的外匯收入中約有超過百分之八來自觀光 收益高居第一，儼然已成為許多國家賺取外匯的重要來源之一。 本研究主要是利用時間數列分析與預測模式等多變量分析方法， 並配合SAS等統計軟體加以建立來台旅客預測之建構模式、預測 及檢定，並據以推論其結果、關係及影響，依據中華民國交通部 觀光局來華人數統計1997年1月至2005年9月的觀光資料統計之結 果了解來華觀光旅客變動情形，我們由研究結果可看出實際和預 測來台旅客人數發展趨勢，發現干預模型比ARIMA有更好的預測 效果。藉由掌握來華觀光旅客旅遊趨勢及型態，建立來華觀光旅 客人數需求。希研究結果提供政府相關機關或地方民間企業之相 關規劃部門，做為對未來我國觀光需求政策方針的擬定規劃之參 考。

In twenty-first Century tourism industry has become one the most important economic indicators of the industry. According to the World Tourism Organization (WTO) report it pointed out that about more than eight percent from tourism which is the highest income in the global foreign exchange earnings for the country, just like it (tourism revenue) has become the primary source of many countries to earn profit from the foreign exchange. This study is the use of multivariate analysis, time series analysis and forecasting methods, and with SAS statistical software, such as the establishment of Taiwan visitors to construct predictive model to predict and test. In addition, according to the statistics the number of tourist statistics from January 1997 to September 2005 to Taiwan visitors provided by Tourism Bureau, Republic of China Ministry of Transportation. From the inference result, relationship and influence, understand the visitor changes in the situation from time to time, and get to travel trends and patterns of travel to Taiwan visitors. We figure out the actual and forecast trends, the intervention model for better prediction of the number of Taiwan ARIMA. We hope the results can provide the relevant government authorities or the relevant planning department of the local private sector, create to Taiwan visitor number of demands and as demand for the future of tourism policy guidelines of reference drawn planning.