The semantic optimisation of collaborative filtering systems
by Youcef Dahmani; Lalia Benothmane; Sahraoui Kharroubi
International Journal of Metadata, Semantics and Ontologies (IJMSO), Vol. 10, No. 2, 2015

Abstract: The information filtering is the process for routing a continuous flow of relevant information towards groups of people, without having to explicitly express what they want. This process provides ranked lists of recommendation of items to be of use to a user. This system collects from users their ratings, evaluations or preferences for certain items. We distinguish different classes of recommendation approaches, in this work we focus on collaborative filtering. The development of such a system suffers from problems, such as the initial construction phase 'cold start', the scalability and data sparsity. In this paper, we combine the ontology 'item' of semantic web with the memory-based algorithm used in information filtering to reduce some drawbacks of collaborative filtering. We enrich the system by a new term which is 'representative' to fix the problem of funnel effect. We propose a new function which calculates the similarity between users in case of lack of items' evaluations and furthermore, we suggest a new algorithm to reduce the problem of sparse matrix 'item/users'.

Online publication date: Tue, 28-Jul-2015

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Metadata, Semantics and Ontologies (IJMSO):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com