Aspect rating analysis based product ranking

Abstract

Nowadays, the consumer reviews for various products are playing a very important role not only for consumers but also for the firms. A large collection of consumer reviews is now available on the internet. These reviews are very helpful to get quality information about the products. The consumer reviews are used as a feedback by the firms in their product development strategies and consumer relationship management. The consumer reviews contain valuable information still we face difficulties in information navigation due to their disorganized nature. The existing product aspect ranking framework automatically identifies important aspects of products from consumer reviews. There are two important observations to identify important aspects. The large number of consumers usually comments important aspects of the product and the consumers' opinion on those aspects have a great influence on their overall opinion about the product. It uses shallow dependency parser for identifying product aspects and sentiment classifier for determining opinion on those aspects. Finally, it uses a probability aspect ranking algorithm to infer the importance of aspects and ranks it as per their importance score. In this paper, the experimental results confirms the proposed modified system makes the use of aspect rating to improve the performance of important aspect identification and ranking.