Semantic User Profiling for News Domain

Online news reading is becoming an important part of the daily routine of online web users. The amount of information available online is increasing exponentially. Although this information is a valuable resource but lots of scattered, unstructured, irrelevant data is increasing on the web every moment which limits its value. Many research projects and companies are exploring the use of personalized applications that manage this deluge by tailoring the information presented to individual users. These applications all need to gather, and exploit, some information about individuals in order to improve the relevance of news recommended to the end user. User behavior analysis is very important step of news recommendation. As a very first step we have identified different problems faced by online news readers and problems faced in user profiling of online news readers. Then in our proposed approach we are trying to curb these problems to make the system more efficient for the end user. For semantic user profiling we have designed ontologies for various categories of news items. Knowledge of news domain captured in ontologies, act as a classifier for news items and help in semantic user profiling as well. We are incorporating International Press Telecommunication Council (IPTC) standards in our design since IPTC has proposed various standards to make the system more interoperable. We have classified RSS (Really Simple Syndication) feed news items into categories specified in our designed ontologies. Characteristics and preferences of online news readers have been analyzed semantically. In user profiling for news domain we are considering properties of news items also, along with user information. Information has been collected explicitly, through direct user intervention, as well as implicitly, through agents that monitor user activity. This user profiling helps us to better understand our end user. Ontological inference in making user profiling also helps to find the user interests in the area which could not be identified from the user browsing history directly. In contrast to the previous work done in the area, our system can handle issues mentioned in the paper in a better way.