An empirical examination of consumer behavior for search and experience goods in sentiment analysis

With the explosive increase of user-generated content such as product reviews and social media, sentiment analysis has emerged as an area of interest. Sentiment analysis is a useful method to analyze product reviews, and product feature extraction is an important task in sentiment analysis, during which one identifies features of products from reviews. Product features are categorized by product type, such as search goods or experience goods, and their characteristics are totally different. Thus, we examine whether the classification performance differs by product type. The findings show that the optimal threshold varies by product type, and simply decreasing the threshold to cover many features does not guarantee improvement of the classification performance.