本研究利用層級貝氏分析法推估，以解決由聯合分析所導致引起的兩難困境，聯合分析可以輕易地從受測者的人口屬性判斷出遊程的偏好，然而受測者於喜好排序時，必須面對由各種屬性所組成的大量遊程商品。本研究的目的是，在產生候選遊程商品時，各種屬性於組成遊程商品時僅使用一次，減少候選遊程以利受測者進行喜好排序，並且採用階層貝氏層級分析法隨機建立關係域，藉以利用受測者的人口屬性推估其遊程的偏好，且得到與傳統方法相近的準確推估。
研究中共完成55份初測問卷，並於臺北市、臺中市及高雄市三個臺灣主要都會區完成含括本國籍369份及與非本國籍人士90份共459份之問卷實測。受訪者對於遊程選擇包括旅遊的動機、旅遊型態、旅遊費用、旅遊伴侶。
本研究作為決策推薦系統的先期研究，成功地根據遊客的人口統計特性區隔台南市溫泉旅遊市場，一旦遊客和遊程之間的關係清楚地確定，未來可有效地應用於台南市溫泉旅發展的推廣。 This study adapted the hierarchical Bayesian estimation to solve the dilemma commonly induced in the conjoint analysis. The conjoint analysis can easily determine the itinerary preference according to the subjects’ demographic attributes. The subjects always face the trade-off problem in sorting overburden itinerary products composed of various features, because the combination of various features can generate overburden itinerary products for sorting. The object of this study is to reduce the candidate itineraries by using all possible features once in composing all itinerary products for sorting, and to adopt the hierarchical Bayesian estimation to randomly construct the relationship domain to interpret the itinerary preference according to the subjects’ demographic attributes without losing as much estimation accuracy as in the traditional
conjoint analysis.
The survey implemented 55 questionnaires for the initial test, and the final survey has carried out total 459 effective questionnaires with 369 natives and 90 foreigners in three major cities, including Taipei, Taichung, and Kaohsiung, in Taiwan. The features composing the candidate itineraries include travelers’ motive, travel style, travel fee, and travel companion.
As the preliminary study of the decision-support system, this study successfully delineated the market segmentation for hot spring tourism in Tainan City according to travelers’ demographic attributes. Once the relationship between the travelers and the itineraries can be clearly identified, the further promotion for hot spring tourism in Tainan City can effectively carried out.
