A Method for Discovering Local Travel Information from Travel-blogs

Information mining for travel has become a popular research domain in the development of tourism industry nowadays. Currently, studies focusing on the extraction of available travel information from actual tourists' evaluations and reviews on social media sites have been conducted under numerous occasions. In this study, we have proposed another effective yet simple method of discovering local travel-information from traveler's blogs by locating region-sensitive information for the travelers. In this process, regionally restricted words are first extracted based on their frequency of appearances inside the blog. Then, a region-restrictedness score of each blog is calculated through the analysis of the previously extracted regionally restricted phrases. From there, we begin analyzing the content of each blog and classify them into pre-defined categories, using LDA model and word embedding-representations. Through this process, we are able to generate blog recommendations based on our method and see some successful examples as a proof for the effectiveness of our approach.