Cluster-Smoothed with Random Neighbor Selection for Collaborative Filtering

Collaborative filtering is an approach that is usually used for recommendation system to get prediction value from item by user active. Sometimes user not fully gives rating toward all items that caused the rating data becomes sparse. In Collaborative filtering, for handling this problem we can do smoothing process. This paper implemented Cluster-Smoothed method as smoothing process and used Random Neighbor Selection method for determining neighbor that helps in prediction process. Based on research, the smallest Mean Absolute Error (MAE) value obtained is 0.732.