Automatic Annotation of Non-English Web Content

Abstract

Nowadays we are facing the daily information overload. It is thus difficult to get exactly the information we need. It often happens that while reading, we find a word we do not understand and we would need an explanation or some additional information about this word. For this purpose annotations in the Web environment are created and attached to such words. In this paper we propose a method for an automatic extension of the content available on the Web by adding annotations to selected terms (keywords) in the text. The method is designed to be able to insert annotations into the text written in Slovak with a potential to be language independent. Annotations themselves are obtained through publicly available services providing information retrieval. We adapt created annotations taking into account implicit feedback from users in form of click through data. We evaluate the proposed method in the environment of an educational web-based system.