Utilizing Social Trust Mechanisms to Minimize the Influence of False Ratings on Recommendation Systems
Sin-Hua Wu, Lien-Fa Lin · 2024
In today's recommendation systems, people's ratings of products are important criteria for analysis. For instance, a recommendation system based on content analysis may be utilized. However, this can lead to a decrease in recommendation accuracy due to the presence of a large number of false ratings. In this study, we propose a recommendation system that utilizes two different neural networks. The first model is based on users' ratings of items for rating prediction, while the second model incorporates users' social trust data to determine the trust level of item ratings. Finally, the results of the two models are integrated using an integrated learning approach to mitigate the effect of false ratings.