Personalizing Education News Articles Using Interest Term and Category Based Recommender Approaches
Akhilan Subramanian · International Journal of Computer Science and business informatics/International journal of computer science and business informatics · 2014
Speedy growth of internet technologies has provided sufficient ways and mechanisms to provide any information to any person about any entity irrespective of time and place. Getting knowing what had happened around an individual in their living place is an essential requirement in this modern world. This is achieved with the help of various news service providers such as yahoo, google, your news etc. While delivering news to its users most of the news service providers do not take into the account the user’s choice or interests for information content delivery. Providing all news to all will not be an appropriate one when there exists different class of viewers. This major problem can be solved by adopting personalization and it is the key factor which aims at providing the appropriate data for the related person. It is most essential to personalize the news documents such that users can be made comfort by delivering the news of their preferences or interests. When recommending news items, most of the traditional algorithms are based on TF-IDF, i.e., a term-based weighting method which is mostly used in information retrieval and text mining. However, many new technologies have been made available since the introduction of TF-IDF. This paper proposes a new method for recommending news items based on Extended TF-IDF (ETF-IDF) and Modified Concept Frequency - Inverse Document Frequency (MCF-IDF). ETF-IDF with the semantics of domain ontology, resulting in Modified Concept Frequency - Inverse Document Frequency (CF-IDF) which yields better results than extended term based classification method. MCF-IDF is built and tested in Athena, a recommender extension to the Hermes Genesis News Portal Platform. Athena employs a user profile to store concepts or terms found in news items viewed and browsed by the user. The framework recommends new articles to the user using ETF-IDF recommender and the MCFIDF recommender. A statistical evaluation of both methods shows that the use of ontology significantly improves the performance of a traditional recommender.