A ROle-Oriented Filtering (ROOF) approach for collaborative recommendation. Issue 7 (1st September 2016)

• Record Type: Journal Article
• Authors: Ghani, Imran; Jeong, Seung Ryul
• Is Part Of: Enterprise information systems. Volume 10:Issue 7(2016)
• Page Start: 697
• Page End: 728
• Publication Date: 2016-09-01
• Subjects: role-oriented collaborative filtering -- recommendation -- user profile -- ontology -- ROOF; Information storage and retrieval systems -- Periodicals; Management information systems -- Periodicals; Electronic commerce -- Periodicals; 658.4038011
• DOI: 10.1080/17517575.2014.986213
• Languages: English
• ISSNs: 1751-7575

Abstract

Abstract : In collaborative filtering (CF) recommender systems, existing techniques frequently focus on determining similarities among users' historical interests. This generally refers to situations in which each user normally plays a single role and his/her taste remains consistent over the long term. However, we note that existing techniques have not been significantly employed in a role-oriented context. This is especially so in situations where users may change their roles over time or play multiple roles simultaneously, while still expecting to access relevant information resources accordingly. Such systems include enterprise architecture management systems, e-commerce sites or journal management systems. In scenarios involving existing techniques, each user needs to build up very different profiles (preferences and interests) based on multiple roles which change over time. Should this not occur to a satisfactory degree, their previous information will either be lost or not utilised at all. To limit the occurrence of such issues, we propose a ROle-Oriented Filtering (ROOF) approach focusing on the manner in which multiple user profiles are obtained and maintained over time. We conducted a number of experiments using an enterprise architecture management scenario. In so doing, we observed that the ROOF approach performs better in comparison with other existing collaborative filtering-based techniques.