Collaborative Filtering with Entropy-Driven User Similarity in Recommender Systems

Abstract

Collaborative filtering (CF) is the most popular approach in personalized recommender systems. Although CF approaches have successfully been used and have the advantage in that it is unnecessary to analyze item content when generating recommendations, they nevertheless suffer from problems with accuracy. In this paper, we propose a new CF approach to improve recommendation performance. First, a new information entropy‐driven user similarity measure model is proposed to measure the relative difference between ratings. A Manhattan distance‐based model is then developed to address the fat tail problem by estimating the alternative active user average rating. The effectiveness of the proposed approach is analyzed on public and private data sets. As a result of the introduction of the new similarity measure and average rating estimation, we demonstrate that the proposed new CF recommendation approach provides better recommendations.

A multi-level collaborative filtering method that improves recommendations

“…The difference among the ratings of the items that users have co-rated is the value of similarity that exists between these users. Many similarity measures get better accuracy by manipulating how these calculations take place and the methods are based on absolute rating differences (Wang et al, 2015). According to this approach, it is crucial that all of the values of corated items between two users must be the same.…”

A dynamic multi-level collaborative filtering method for improved recommendations

“…Our proposed method differs with the aforementioned methods since it is based on PCC and adds positive or negative adjustments based on information derived from PCC, such as the similarity value and the number of co-rated items. PCC and Cosine are the most widely used methods by collaborative filtering to provide recommendations [14,[28][29][30]. Our main objective is to adjust the similarity value provided by PCC, either in a positive or in a negative way.…”

“…Other similarity approaches include the one proposed in [13] a fuzzy method is used to assign different weights to different sets of ratings. In [14] an entropy-based method is proposed, where the similarity values are calculated using a collaborative filtering function modified to use entropy. Furthermore, in [15] a similarity method is proposed that takes into consideration both the local and global user behaviour.…”

A secured context-aware tourism recommender system using artificial bee colony and simulated annealing

“…In addition, their work identified and critically evaluated recent significant works in contextual recommendation. Wang et al (2015) proposed an efficient collaborative filtering algorithm for similar users. The algorithm deploys entropy-driven model to compute users' similarities and Manhattan distance-based model to perform rating and recommendation.…”