With the convergence of social networks and recommender systems, new algorithms have been explored, that combine observations on item selection with an analysis of social relations among people. However, they focus on direct friendship relations. In this paper, concerning collaborative filtering approaches, we test the impact of integrating a measure of common friendship, in order to capture the intuition that socially related groups of people tend to have similar tastes. An experiment on the Yelp dataset shows that using preference information derived from the commonalities of interests in networks of friends achieves higher accuracy than item-to-item and other collaborative filtering algorithms.
Enhancing Collaborative Filtering with Friendship Information
ARDISSONO, Liliana;PETRONE, GIOVANNA;SEGNAN, MARINO
2017-01-01
Abstract
With the convergence of social networks and recommender systems, new algorithms have been explored, that combine observations on item selection with an analysis of social relations among people. However, they focus on direct friendship relations. In this paper, concerning collaborative filtering approaches, we test the impact of integrating a measure of common friendship, in order to capture the intuition that socially related groups of people tend to have similar tastes. An experiment on the Yelp dataset shows that using preference information derived from the commonalities of interests in networks of friends achieves higher accuracy than item-to-item and other collaborative filtering algorithms.File | Dimensione | Formato | |
---|---|---|---|
umap2017-2pages.pdf
Accesso aperto
Descrizione: Articolo principale
Tipo di file:
POSTPRINT (VERSIONE FINALE DELL’AUTORE)
Dimensione
506.79 kB
Formato
Adobe PDF
|
506.79 kB | Adobe PDF | Visualizza/Apri |
UMAP-2017-extAbs-PDFEditoriale.pdf
Accesso riservato
Descrizione: Articolo principale
Tipo di file:
PDF EDITORIALE
Dimensione
933.26 kB
Formato
Adobe PDF
|
933.26 kB | Adobe PDF | Visualizza/Apri Richiedi una copia |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.