Abstract
Recently, recommendation methods based on the similarity between different users or objects have achieved remarkable success. However, the high similarity between users does not represent a similar preference in reality. In fact, what really reflects user preferences is user's subjective evaluation on items. In this paper, we propose a method to predict users' evaluations of films by using random forest regression model. As the user's evaluation on items depends on the characteristics of items and preferences reflected in history, we utilize the above two data as inputs to predict users' evaluations of films based on users' rating process simulated by random principle. The results show that the method proposed in this paper outperforms others in MAE.
Recommended Citation
Ma, Yingxue and Gan, Mingxin, "A Random Forest Regression-based Personalized Recommendation Method" (2018). PACIS 2018 Proceedings. 170.
https://aisel.aisnet.org/pacis2018/170