Personal agent for highlighting interesting words

We present a personal intelligent agent which automatically highlights the words that are interesting to a particular user while he/she is browsing web pages. The visited web page is first classified into one of several pre-defined categories according to its content. Since each category is described by a set of keywords, which are user-dependent and detected by a statistical learning approach, the agent can highlight the page using these interesting keywords for fast browsing. Our experiments show that this personalized way of highlighting words in a web page enhances the user´s browsing experience.