Knowledge Mining From Travelogues Based On Probabilistic Topic Models

With the prosperity of travel and Web 2.0 technologies, more and more people have willingness to share their travel experiences on the Web (e.g., weblogs, forums, or Web 2.0 communities) in the form of textual travelogues. These travelogues contain rich information, particularly including location-representative knowledge such as attractions, styles and activities. Such knowledge can greatly facilitate other tourists'trip planning, if it can be correctly extracted and summarized. However, since most travelogues are noisy and lack of explicit destination recommendation or visual information, it is difficult for common users to utilize such knowledge effectively. In this paper, to mine location-representative knowledge from a large collection of travelogues, we propose a probabilistic topic model, named as Location-Topic (LT) model. This model has the advantages of (1) differentiability between two kinds of topics, i.e., local topics which characterize locations and global topics which correspond to common themes co-occurring with various locations in travelogues, and (2) representation of locations in the local topic space to encode both location-representative knowledge and pairwise similarity between locations. Some novel applications are developed based on the proposed model, including (1) destination recommendation under either of two criteria, i.e., being similar to a given destination or being relevant to a given travel intention, (2) destination summarization with representative tags and relevant snippets, and (3) travelogue visualization with images depicting the highlights in the text. Based on two large collections of travelogues, the proposed framework is evaluated using both objective and subjective approaches and shows promising results.