ASSR: An Adjustable Scenic Spot Rating System Based on Travel Note Mining

Scenic spot rating is a helpful and straightforward metric to evaluate scenic spots and assist travel plan making. However, existing ratings from the tourism websites are always unadjustable and strongly correlated with the popularity of the spots while different tourists may have different tastes. To solve this problem, we propose an Adjustable Scenic Spot Rating System (ASSR) based on travel note mining. With the help of the information mined from numerous travel notes, which contains the sentiment of tourists towards these numerous spots, our system is able to evaluate the scenic spots properly. We implement sentiment analysis to detect sentiment polarity of the travel notes. Based on this, we bring up a Spot-Topic LDA (ST-LDA) to generate a sentiment score for each spot which reflects its quality. In this case, the popularity of spots becomes an adjustable factor to satisfy the demands of different people. Our model is validated on public databases and has been embedded in the Smart Tourism Services Platform (STSP).