Choice of metrics used in collaborative filtering and their impact on recommender systems

Abstract

Abstract-The capacity of recommender systems to make correct predictions is essentially determined by the quality and suitability of the collaborative filtering that implements them. The common memory-based metrics are Pearson correlation and cosine, however, their use is not always the most appropriate or sufficiently justified. In this paper, we analyze these two metrics together with the less common mean squared difference (MSD) to discover their advantages and drawbacks in very important aspects such as the impact when introducing different values of k-neighborhoods, minimization of the MAE error, capacity to carry out a sufficient number of predictions, percentage of correct and incorrect predictions and behavior when attempting to recommend the n-best items. The paper lists the results and practical conclusions that have been obtained after carrying out a comparative study of the metrics based on 135 experiments on the MovieLens database of 100,000 ratios.

Extended feature combination model for recommendations in location-based mobile services

“…In addition, several distance metrics have been proposed in the literature such as Graph Distance, Common Neighbors, Jaccard Coefficient, Pearson Correlation and Adamic & Adar [13]. The impact of some of them on recommendation systems has been discussed in [6] and [31] as well. In this study, User-User similarity matrix has been constructed by using the friendship network among the users.…”

“…Recall = T P T P + F N (19) Additionally, F-measure as a combination of precision and recall [31] can be computed as shown in Eq. (20).…”

A new collaborative filtering metric that improves the behavior of recommender systems

“…Despite the deficiencies of Pearson correlation, this similarity measure presents the best prediction and recommendation results in CF-based RS [15,16,31,7,35], furthermore, it is the most commonly used, and therefore, any alternative metric proposed must improve its results.…”

“…Memory-based methods [22,37,35,40] use similarity metrics and act directly on the ratio matrix that contains the ratings of all users who have expressed their preferences on the collaborative service; these metrics mathematically express a distance between two users based on each of their ratios. Model-based methods [1] use the ratio matrix to create a model from which the sets of similar users will be established.…”

Hybrid Collaborative Movie Recommender System Using Clustering and Bat Optimization

“…Implicit was derived from user behavior such as navigation and browsing history. Recommender systems can be mainly classified in to content-based, collaborative and hybrid recommender filtering techniques [3].Collaborative filtering (CF) is the most traditional and commonly used approach to generate recommendations. Collaborative-filtering [4] identify neighboring users who have same interests and preferences in the past by calculating similarities between their profiles and recommend the products by computing a weighted average of these neighbor users' rating.…”