Collaborative Filtering System
Salim İldere · 2019 4th International Conference on Computer Science and Engineering (UBMK) · 2019
Recommendation systems using machine learning created. Recommendation systems can be gathered around 3 main topics. Content-based, collaborative filtering and hybrid suggestion systems. Collaborative filtering system is a system where suggestions are made using similar users. In this article, the effects of the change of hidden factor number on error metrics for collaborative suggestion system are examined. In addition, the learning rate of error metrics was examined by keeping the number of hidden factors constant. The proposed suggestion system has 40 hidden factors and the most successful result is obtained by MAE error metric. MSE error metric, the most successful learning rate graph was obtained. The results showing the data of the suggestion system during training are explained with graphics.