Current recommender systems are mainly based on customers' personal information and online behavior. We find that those systems lack efficiency and accuracy. At the same time, we observe the large amount of review data with exponential growth. Based on this observation, we propose a recommender system based on opinion mining. With text mining method we extract the opinion related information from the massive reviews. We analyse the linguistic information and design a two-layer selection algorithm to find the most suitable products for customers. The experiment shows our method has great accuracy, fleasibility, and reliablity.
展开▼
