Research on fuzzy intelligent recommendation system based on consumer online reviews

Abstract

Many consumers only have fuzzy requirement for products, because they are not the experts of the domain who have much experience for products. The system mines explicit attributes and implicit attributes of products from online reviews. Through using semantic analysis technology and building the fuzzy inference rules based on these products attributes, the system can understand the sentiment of the consumers' review which shows system's intelligence. The sentiment words of implicit product attributes are expressed by the fuzzy function, which is the foundation of the sentiment calculation. Finally the experiment proves that our recommendation method is effective and the system can satisfy consumers' requirement. © 2013 Springer-Verlag Berlin Heidelberg.