出版社：Academy & Industry Research Collaboration Center (AIRCC)
摘要：The project proposes to build a system which is capable of extracting business intelligence for amanufacturer, from online product reviews. For a particular product, it extracts a list of the discussedfeaturesand their associated sentiment scores. Online products reviews and review characteristics areextracted from www.Amazon.com.A two level filtering approach is adapted to choose a set of reviews thatare perceived to be useful by customers. The filtering process is based on the concept that the reviewergenerated textual content and other characteristics of the review, influence peer customers in makingpurchasing choices. The filtered reviews are then processed to obtain a relative sentiment score associatedwith each feature of the product that has been discussed in these reviews. Based on these scores, thecustomer's impression of each feature of the product can be judged and used for the manufacturers benefit.
