Abstract:
The importance of effectively categorizing news information is rising with the expansion of digital media in Bangladesh. The problem of sorting NewsML headlines into six individual categories Entertainment, Life, Business, Education, World, and Sports is addressed in this work. The goal is to determine how well six machine learning algorithms—Multi, KNN, DT, RF, LR, and SVC—perform in automated classification. The study aims to improve the performance of the digital news platform user experience and content management system by developing a solid framework for Bangla news classification. The dataset comprises five thousand headlines from prominent NewsML portals, including Prothom Alo, Kaler Kontho, BDNews, and News24. The research shows that the SVC algorithm obtains the best accuracy of 0.99% on a 20% test dataset after extensive preprocessing and analysis. This finding demonstrates how well SVC works when classifying news headlines and dealing with the subtleties of Bangla text.
