Currently, customers’ trading activities are supported by recommender tools that are able to generate personalized suggestions. Many of these recommenders are centralized, lacking in efficiency and scalability, while others are distributed and, conversely, imply a computational overhead on the client side being often excessive or even unacceptable for many devices. In this article, we propose a distributed recommender, based on a multi-tiered agent system, where the agents of each tier are specialized in a different e-commerce activity. The proposed system is capable of generating effective suggestions without a too onerous computational burden. In particular, we show that our system introduces significant advantages in terms of openness, privacy, and security.

Introducing Specialization in e-Commerce Recommender Systems

ROSACI D;SARNE' G
2013

Abstract

Currently, customers’ trading activities are supported by recommender tools that are able to generate personalized suggestions. Many of these recommenders are centralized, lacking in efficiency and scalability, while others are distributed and, conversely, imply a computational overhead on the client side being often excessive or even unacceptable for many devices. In this article, we propose a distributed recommender, based on a multi-tiered agent system, where the agents of each tier are specialized in a different e-commerce activity. The proposed system is capable of generating effective suggestions without a too onerous computational burden. In particular, we show that our system introduces significant advantages in terms of openness, privacy, and security.
File in questo prodotto:
File Dimensione Formato  
palopoli_2013_CERA_introducing_editor.pdf

non disponibili

Tipologia: Versione Editoriale (PDF)
Licenza: Tutti i diritti riservati (All rights reserved)
Dimensione 305.06 kB
Formato Adobe PDF
305.06 kB Adobe PDF   Visualizza/Apri   Richiedi una copia
palopoli_2013_CERA_introducing_post.pdf

accesso aperto

Descrizione: versione postprint
Tipologia: Documento in Post-print
Licenza: Creative commons
Dimensione 3.69 MB
Formato Adobe PDF
3.69 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/20.500.12318/1369
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 17
  • ???jsp.display-item.citation.isi??? 10
social impact