| Summary: | 碩士 === 國立臺東大學 === 文化資源與休閒產業學系 === 102 === Internet platform provides consumers a place to express their ideas and experience. More and more consumers are used to searching information from internet before they shop, thus the influence of electronic word of mouth(e-WOM) is also increasing day by day. Travel market might be effected by e-WOM, climate, environment, even economic variation. Therefore, e-WOM is a appropriate method to discuss the tourists' true feeling in tourism industry.
In this study, blog data were collected from a total of 11,745 articles from 2009 to 2013 about Kenting National Park’s tourism from the Google website. Based on the literature review, the emotion degree of blog posts were acquired by trait words which were chose by scholar and operator in tourism and converted to emotion scores. Furthermore, Google Trends and five tourism-related economic indicators were taken as independent variables. The back-propagation network (BPN) of artificial neural network was used to construct the prediction models of different types of variables combinations to estimate the tourist volume.
In consideration of predicting ability, the emotion scores, Google Trends and five tourism-related economic indicators were shifted forward for 1 to 3 months. The prediction models were evaluated by MAPE. The empirical results show: 1. Taking advantage of electronic word-of-mouth, Google Trends combined with economic indicators can really be used in predicting the travel market. 2. The prediction models that combines electronic word-of-mouth or Google Trends with economic indicators can achieve in a good range of MAPE(10%-20%). 3. The best predicting ability is when emotion scores, Google Trends, and economic indicators were shifted forward for 3 months, indicating that the influential time point for Internet emotion to predict the future volume of tourism in the Kenting National Park travel market is 3 months. This study provides an objective, innovative and convenient way of predicting. It can be used practically as a predicting reference for operators when preparing plans of future operations.
|
|---|
