Collaborative Filtering: Inference from Interactive Web

Collaborative Filtering: Inference from Interactive Web

Tania Al. Kerkiri (University Of Macedonia, Greece) and Dimitris Konetas (University of Ioannina, Greece)
DOI: 10.4018/978-1-61520-921-7.ch012
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

The interactive tools like blogs, wikis, et cetera, known under the commonly acceptable name Web2.0, led to a new generation of Internet services and applications such as social networks, recommendation systems, reputation systems, et cetera, allowing for public participation in the formation of the content of the Web, and at the same time fueling an explosion of information. This information is a widely available intellectual capital. Due to the opportunities that arise from the exploitation of this information, this chapter will i) present the rationale under which these systems function, ii) summarize and apply in an indicative manner the mathematical models used to handle this information, iii) propose a general architecture of these systems and iv) describe a hybrid multifaceted algorithm that exploits the capabilities arising from this information towards a personalized inference for a specific user. The result of this work is an indication of the capabilities that arise from further exploitation of these systems.
Chapter Preview
Top

2. Collaborative-Filtering Methods: What Are They?

Balabanović (1997) declares collaborative-based filtering methods as algorithms that trace items for the user who interacts with the system at any given time, based on information collected from users which have experienced the system up to now (Resnick, 2000). The idea underlying the operation of these systems is based on two basic assumptions:

Complete Chapter List

Search this Book:
Reset