Abstract: We investigate the key determinants of user ratings in social media-based business rating services. Our hypothesis is that beyond factors internal to a user, external factors, such as the direct and indirect influence of other users, and environmental factors beyond the control of the user have a significant role in determining the actual rating assigned by a user for a given service. To test this hypothesis, we used data from Yelp, and attempted to predict user ratings on location-based services, in particular food and restaurant business. Our results show improved prediction performance over the baseline, with improved robustness to rating variability and rating sparsity.
Loading