An Empirical approach for pruningaspects inproduct review datasets using Content Validity for identifying relevant candidate aspects

Abstract

Due to the large amount of opinions available on the websites, consumers are often overwhelmed with information and find it extremely difficult to use the available information to make a decision about the products they purchase.Opinions of these consumers are classified as expressing Positive, Negative or Neutral sentiments about the aspects of the products. From the reviews of the consumers, Candidate Aspects are extracted using feature extraction methods. Later, Aspect Pruning is performed in order to remove the aspects which are not relevant to the product. Finally, theIdentified Candidate aspects are validated by the experts using Content Validity. This paper validates the candidate aspects of 5 product (Apex AD2600 Progressive-scan DVD player, Canon G3, Creative Labs Nomad Jukebox Zen Xtra 40GB, Nikon coolpix 4300 and Nokia 6610)datasets.