Constructing Predictive Models and Investigating Main Factors for Repeat Visitation to Taiwan

• Main Authors: Hsin-Lan Lin, 林欣嵐
• Format: Others
• Language: en_US
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/49992849541437574587

碩士 === 國立臺中技術學院 === 事業經營研究所 === 95 === Global tourism has become one of the most potent industries in the world in recent years. Simiarly, tourism has been gaining popularity in past years in Taiwan. After the effect of Severe Acute Respiratory Syndrome (SARS), the Taiwan government started to open the door of Taiwan tourism to the people from mainland China and held many activities to attract tourists. After the review of literature, it is costly to capture new customers from competitors because a greater degree of service improvement is needed to convince customers to switch from competitors. Hence, methods that can forecast repeat visitation accurately are greatly needed. The purposes of this study are to build predictive models and find the main factors for repeat visitation. In addition this study mainly investigates the relationship amng travel motivation, tourist satisfaction and intention of repeat visitation. A total of 307 questionaires were found valid and the findings of this study are summarized as follows. First of all, the results indicated that the main motivations of traveling in Taiwna are “Modernity and infrastructure,” “Climate and surroundings,” “Differebt cultures,” “Nightlife and entertainment,” “Education and knowledge,” “Relaxation and escaping,” Exciting” and “Togetherness.” In addition, some siginificant differences were foumd between demographic characteristics of tourists and the three construct, travel motivations, tourist satisfaction, and repeat visitation. Moreover, the positive relationship among travel motivation, tourist satisfaction and repeat visitation was identified. That explained travel motivations drive people to travel. With travel experience, tourists have some levels of satisfaction, which can directly lead to the intention of repeat visitation. Finally, three predictive models were constructed by Binary Logistic Regression, Support Vector Machines and Clustering-Launched Clssification. The three models possessed good predictive ability.