Academic journal article Journal of Digital Information Management

The NEWER System: How to Exploit a Neuro-Fuzzy Strategy for Web Recommendation

Academic journal article Journal of Digital Information Management

The NEWER System: How to Exploit a Neuro-Fuzzy Strategy for Web Recommendation

Article excerpt

1. Introduction

During the past few years the World Wide Web has become the biggest and the most popular way of communication and information dissemination. It serves as a platform for exchanging various kinds of information, ranging from research papers, and educational content, to multimedia content, software and personal blogs. Everyday, the Web grows by roughly millions of electronic pages, adding to the hundreds of millions pages already on-line. Because of its rapid and chaotic growth, the resulting network of information lacks of organization and the structure of Web sites becomes more and more complex. As a consequence, when browsing and searching the Web, users are very often overwhelmed by a huge amount of information available online and get lost in that information overload problem that continues to expand. These factors combined with the heterogeneous nature of the Web make the browsing activity difficult not only for ordinary users, who often are faced with the challenging problem of finding the desired information in right time.

An important step in the direction of alleviating the problem of information overload is represented by Web personalization.

Web personalization can be simply defined as the task of adapting the information or services provided by a Web site to the needs and interests of users, exploiting the knowledge gained from the users' navigational behavior and individual interests, in combination with the content and the structure of the Web site (Mobasher, 2005).

Within the field of Web personalization, Web recommendation is a promising technology aimed to predict the user interests, by providing them with the information or services that they need without explicitly asking for them (Mulvenna et al., 2000), (Nasraoui, 2005).

Generally speaking, in the development of a Web recommendation system, two main problems have to be addressed:

* how to discover knowledge about user interests from the large amount of Web data;

* how to exploit the discovered knowledge in order to deliver intelligent recommendations to new users.

Both these tasks are characterized by uncertainty and vagueness. In effect, a great deal of ambiguity pervades all stages of the user interactions: from the definition of the user navigational models to the recommendation process. Indeed, Web data are uncertain and not accurate usually because of partial Web information. To deal with these properties, the traditional and hard computing methods revealed to be inadequate and they require to be extended and enhanced.

Due to their characteristics, Soft Computing (SC) techniques seem to be particularly appropriate tools that allow to accommodate the needs of the Web by overcoming the limitations of hard computing methods (Frias-Martinez et al., 2005). In fact, the role model for SC is the human mind. Hence, unlike the conventional methods, SC reveals to be tolerant of imprecision, uncertainty, partial truth, and approximation. More precisely, SC refers to a consortium of methodologies that work synergistically to find approximate solutions for realworld problems which contain various kinds of inaccuracies and uncertainties. The guiding principle for these paradigms is to devise methods of computation that lead to an acceptable solution at low cost by seeking for an approximate solution to an imprecisely/precisely formulated problem. Computing paradigms underlying SC are neural network, fuzzy logic and genetic algorithms.

Rather than a collection of different paradigms, SC can be better regarded as a partnership in which each of the partners provides a methodology for addressing problems in a different manner.

From this perspective, SC methodologies are complementary rather than competitive. This relationship enables the creation of hybrid computing schemes which use neural networks, fuzzy systems and evolutionary algorithms in combination allowing to overcome limitations and to exploit advantages of each single paradigm (Hildebrand, 2005), (Tsakonas et al. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.