News Category Network based Approach for News Source Recommendations

Abstract

News sources overall distribute their news content utilizing category words in their URLs to keep up the Web directory structure on their sites. These news categories can be of two sorts either generic, for example, 'Business, Sports, etc.' or time-specific 'ICC Champions Trophy 2017' and 'Delhi Earthquake' or both. These news categories can be utilized for different applications, for example, news source recommendations and time specific news category extraction. The current frameworks, for example, Dmoz directory and Yahoo registry are mostly human annotated and do not consider the time dynamism of categories of news sources. Incited by aforementioned issues, in this paper, we propose a system that recommends the news sources with the help of news category network for the given news category. To build news category network, we use a dynamic approach that iteratively modifies the category network on the basis of URL categories. We additionally propose an approach that rank the news sources for the given categories using three parameters such as Traffic Based Website Importance, Social media Analysis and Category Wise Article Freshness. Experiments are done on category URLs captured from GDELT project to evaluate our proposed system.