Ensemble Similarity based Collaborative Filtering Feedback: A Recommender System Scenario

One significant aspect of any data mining associated task is the fundamental requirement of data for knowledge discovery. In most of the cases, the data are available for the desired task which produces recommendations for a user for a given set of items, where items could be movies, books, articles or products. The major issue is that most of the movies are not rated and are untagged. This makes it difficult to judge based on its content and views by community. Due to the insufficiency in data, recommender systems cannot produce recommendations of rare items. The objective of this paper is to improve the performance of recommender systems by making better recommendations using corresponding implicit data and ensemble of similarity metrics to achieve accurate predictions of the user ratings. This paper proposes using both explicit and implicit feedback as data sources for input in recommender systems along with an ensemble of similarity measures to get better and more useful recommendations.