Optimization of the Neighbor Parameter of k-Nearest Neighbor Algorithm for Collaborative Filtering

摘要

Collaborative Filtering (CF) is one of the prime techniques used in the field of Recommender System. The recommender system is used for predicting the preference of the user based on his earlier preferred items. This process of predicting involves k-Nearest Neighbor (kNN) method, to find the users with similar type of preferences, interest or taste. In this paper, the experiments are carried out to check the influence of parameter k on the results obtained from kNN algorithm and to find the value of k for which we get the optimal accuracy for the kNN algorithm.