An intelligent and adaptive news recommender system using modified collaborative filtering for user profiling

Search Engines on the Internet help find web content meeting specific criteria, typically containing a given word or phrase. But they fail to take into account personal interests of the users. Researchers have been working over the years to come up with more adaptive and evolving systems that can help the information retrieval community on the internet to get more individualized service. Recommender systems are getting popular as they suggest information sources and products based on learning user likes and dislikes.; The problem of information overload is particularly relevant to the users looking for online news resources. Existing recommender systems commonly use content-based approach and collaborative-based approach to predict new items of interest to the user. This paper proposes a system that combines both these approaches in an efficient framework leading to a more productive and accurate system.; It implements a modified collaborative filtering algorithm for dynamic user profile modeling that tries to minimize user effort involved in the learning procedure. To determine the overall accuracy and productivity of the proposed system, a set of experiments were performed and an existing similar system is used for comparison purposes. In the end some directions for future work to further refine the system.