In this paper, a search method of tourism information based on attraction dynamic topic distribution is proposed. We constructed the attraction dynamic topic model and trained the model with tourism comments of different time period to get the topic distribution of attractions in time dimension. In this way, we can speculate user's search intents by topic distribution of the query. Based on experiments on tourism comments of Beijing attractions crawled from websites, we verified the dynamic change of attraction topics and the accuracy of our proposed search method. In the future, we plan to consider more of users' personalized factors, and establish a more optimized user search intent model to achieve better travel information search.
展开▼
