Word sense disambiguation of the polish language for the needs of generating digital identities

Abstract

Nowadays, as more and more real world activities are transferred to the Web, the topic of digital identities, understood as representations of users' characteristics, become increasingly important. One of the purposes of research into digital identities is to use them as a means of informing the systems the user interacts with about his/her characteristics, needs and expectations, in order to, for example, enable a constant personalization process. However, a requirement that such identities are to represent a wide range of the user's characteristics in the form understandable to many different systems, raises many difficulties. An issue of how to collect such a broad set of information about users is also of particular significance. The aim of the article is to analyze two topics that can potentially be used while solving the above mentioned problems. These are ontology-based user modeling, which can be used for building reusable user models, and word sense disambiguation for words extracted from articles read by the users on different sites they visit. Based on the analysis of these topics, a proposal of further research directions is presented.