Details

Autor(en) / Beteiligte

• An, Ya-Hui
• Dong, Qiang
• Sun, Chong-Jing
• Nie, Da-Cheng
• Fu, Yan

Titel

Diffusion-like recommendation with enhanced similarity of objects

Ist Teil von

• Physica A, 2016-11, Vol.461, p.708-715

Ort / Verlag

Elsevier B.V

Erscheinungsjahr

2016

Quelle

Access via ScienceDirect (Elsevier)

Beschreibungen/Notizen

• In the last decade, diversity and accuracy have been regarded as two important measures in evaluating a recommendation model. However, a clear concern is that a model focusing excessively on one measure will put the other one at risk, thus it is not easy to greatly improve diversity and accuracy simultaneously. In this paper, we propose to enhance the Resource-Allocation (RA) similarity in resource transfer equations of diffusion-like models, by giving a tunable exponent to the RA similarity, and traversing the value of this exponent to achieve the optimal recommendation results. In this way, we can increase the recommendation scores (allocated resource) of many unpopular objects. Experiments on three benchmark data sets, MovieLens, Netflix and RateYourMusic show that the modified models can yield remarkable performance improvement compared with the original ones. •The distributions of RA similarity follow almost power-law.•The RA similarity is a key factor in measuring the resource transfer between objects.•The distribution of ERA similarity is more even, thus the similarity of a large proportion pairs of unpopular objects is increased.•The enhancement of similarity greatly improves both of accuracy and diversity of diffusion-like models.•The proposed method can be analogously applied to most similarity-based recommendation models on user–object bipartite networks.