Today, Web personalization offers valid tools for the development of applications that have the attractive property to meet in a more effective manner the needs of their users. To do this, Web developers have to address an important challenge concerning the discovery of knowledge about interests that users exhibit during their interactions with Web sites. Web Usage Mining (WUM) is an active research area aimed at the discovery of useful patterns of typical user behaviors by exploiting usage data. Among the different proposed techniques for WUM, clustering has been widely employed in order to categorize users by grouping together users sharing similar interests. In particular, fuzzy clustering reveals to be an approach especially suitable to derive user categories from Web usage data available in log files. Usually, fuzzy clustering is based on the use of distance-based metrics (such as the Euclidean measure) to evaluate similarity between user preferences. However, the use of such measures may lead to ineffective results by identifying user categories that do not capture the semantic information incorporated in the original Web usage data. In particular, in this chapter, we propose an approach based on a relational fuzzy clustering algorithm equipped with a fuzzy similarity measure to derive user categories. As an application example, we apply the proposed approach on usage data extracted from log files of a real Web site. A comparison with the results obtained using the cosine measure is shown to demonstrate the effectiveness of the fuzzy similarity measure.
How to derive fuzzy user categories for Web personalization
CASTELLANO, GIOVANNA;
2009-01-01
Abstract
Today, Web personalization offers valid tools for the development of applications that have the attractive property to meet in a more effective manner the needs of their users. To do this, Web developers have to address an important challenge concerning the discovery of knowledge about interests that users exhibit during their interactions with Web sites. Web Usage Mining (WUM) is an active research area aimed at the discovery of useful patterns of typical user behaviors by exploiting usage data. Among the different proposed techniques for WUM, clustering has been widely employed in order to categorize users by grouping together users sharing similar interests. In particular, fuzzy clustering reveals to be an approach especially suitable to derive user categories from Web usage data available in log files. Usually, fuzzy clustering is based on the use of distance-based metrics (such as the Euclidean measure) to evaluate similarity between user preferences. However, the use of such measures may lead to ineffective results by identifying user categories that do not capture the semantic information incorporated in the original Web usage data. In particular, in this chapter, we propose an approach based on a relational fuzzy clustering algorithm equipped with a fuzzy similarity measure to derive user categories. As an application example, we apply the proposed approach on usage data extracted from log files of a real Web site. A comparison with the results obtained using the cosine measure is shown to demonstrate the effectiveness of the fuzzy similarity measure.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.