Host: The Japanese Society for Artificial Intelligence
Name : The 36th Annual Conference of the Japanese Society for Artificial Intelligence
Number : 36
Location : [in Japanese]
Date : June 14, 2022 - June 17, 2022
How to identify helpful information from large-scale online reviews has become a prominent issue in studies of wisdom-of-crowds. In this research, we focused on the online reviews with heretical opinions (i.e., heretical reviews). We examined whether these reviews could provide more helpful information compared to reviews with orthodox opinions. Using sentiment analysis, sentence-embedding methods, and simulations based on Bayesian inference for a large-scale dataset, we found that heretical reviews were deemed more helpful because they provided more sufficient, neutral, and unique information. To interpret these results, we considered that the reviewers of heretical reviews could face peer pressure when expressing heretical opinions. This peer pressure motivated the reviewers of heretical reviews to offer more convincing evidence (i.e., sufficient, neutral, and unique information) to persuade their readers. This study offers a simple, but effective approach to elicit helpful information from a large-scale of online reviews. Additionally, this study also provided a deeper understanding of the mechanism underlying online review behaviors: previous studies always considered that peer pressure causes biases in collective online behavior; however, this study uncovered that peer pressure can cause valuable outcomes in online review behaviors.
