Hybrid recommendation methods in complex networks

Fiasconaro, A. (Universidad de Zaragoza) ; Tumminello, M. ; Nicosia, V. ; Latora, V. ; Mantegna, R.N.

Resumen: We propose two recommendation methods, based on the appropriate normalization of already existing similarity measures, and on the convex combination of the recommendation scores derived from similarity between users and between objects. We validate the proposed measures on three data sets, and we compare the performance of our methods to other recommendation systems recently proposed in the literature. We show that the proposed similarity measures allow us to attain an improvement of performances of up to 20% with respect to existing nonparametric methods, and that the accuracy of a recommendation can vary widely from one specific bipartite network to another, which suggests that a careful choice of the most suitable method is highly relevant for an effective recommendation on a given system. Finally, we study how an increasing presence of random links in the network affects the recommendation scores, finding that one of the two recommendation algorithms introduced here can systematically outperform the others in noisy data sets.

Idioma: Inglés

DOI: 10.1103/PhysRevE.92.012811

Año: 2015

Publicado en: Physical Review E 92, 1 (2015), 012811 [10 p.]
ISSN: 2470-0045

Factor impacto JCR: 2.252 (2015)

Categ. JCR: PHYSICS, MATHEMATICAL rank: 6 / 53 = 0.113 (2015) - Q1 - T1
Categ. JCR: PHYSICS, FLUIDS & PLASMAS rank: 10 / 30 = 0.333 (2015) - Q2 - T2

Factor impacto SCIMAGO: 1.183 - Condensed Matter Physics (Q1) - Statistical and Nonlinear Physics (Q1) - Statistics and Probability (Q2)

Tipo y forma: Artículo (Versión definitiva)

Área (Departamento): Área Física Materia Condensada (Dpto. Física Materia Condensa.)