Categorizing online news articles using penguin search optimization algorithm

Abstract

Online news is an emerging channel where the internet users can get news. Analyzing huge volume of online news articles is a challenging one, because online news articles are generated and updated time to time. Big data techniques are used to tackle this problem. In order to classify the news articles into different categories, an approach based on Evolving Fuzzy Systems(EFS) was used. It categories news articles based on the changes in the content of the corresponding articles. However, it has the problem in the selection of threshold value. Moreover Gaussian membership function is used in EFS that describes the closeness to the prototype. Sometimes it is ha rd to justify. So in this paper, a Penguins Search Optimization Algorithm(PeSOA) is introduced to optimize the pruning threshold value and a bell shaped fuzzy membership function is introduced to define the closeness to the prototype. The optimized pruning threshold is used in term filtering which prune the generated terms based on their frequencies of occurrence throughout the collection. Then the fuzzy rules are generated by EFS where bell shaped fuzzy membership function is used to define the closeness to the prototype. Based on t he fuzzy rules the online news articles are categorized.