A New Similarity Measure Based on Mean Measure of Divergence for Collaborative Filtering in Sparse Environment

Memory based algorithms, often referred to as similarity based Collaborative Filtering (CF) is one of the most popular and successful approaches to provide service recommendations. It provides automated and personalized suggestions to consumers to select variety of products. Typically, the core of s...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Procedia computer science Jg. 89; S. 450 - 456
Hauptverfasser: Suryakant, Mahara, Tripti
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Elsevier B.V 2016
Schlagworte:
Collaborative Filtering
Information Filtering
Recommendation Accuracy
Recommendation System
Similarity Measure
Recommendation System
Information Filtering
Collaborative Filtering
Similarity Measure
Recommendation Accuracy
ISSN:1877-0509, 1877-0509
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Memory based algorithms, often referred to as similarity based Collaborative Filtering (CF) is one of the most popular and successful approaches to provide service recommendations. It provides automated and personalized suggestions to consumers to select variety of products. Typically, the core of similarity based CF which greatly affect the performance of recommendation system is to finding similar users to a target user. Conventional similarity measures like Cosine, Pearson correlation coefficient, Jaccard similarity suffer from accuracy problem under sparse environment. Hence in this paper, we propose a new similarity approach based on Mean Measure of Divergence that takes rating habits of a user into account. The quality of recommendation of proposed approach is analyzed on benchmark datasets: ML 100K, ML-1M and Each Movie for various sparsity levels. The results depict that the proposed similarity measure outperforms existing measures in terms of prediction accuracy.
ISSN:1877-0509
1877-0509
DOI:10.1016/j.procs.2016.06.099