Automated Regional Language News Article Classification System for Kannada and Telugu

The increasing amount of digital news content has created a pressing need for automated systems that can efficiently organize and categorize articles. This study focuses on developing an automated classification system for Kannada and Telugu news articles, targeting predefined categories. For Kannada, the categories include Entertainment, Sports, and Lifestyle, while for Telugu, the categories are Sports, Business, and Entertainment. With the help of the techniques available in natural language processing and the available models in machine learning, the system preprocesses the textual data in both languages to extract meaningful features for classification. Several classification algorithms are explored, with a focus on evaluating model performance using standard metrics. The results demonstrate that embeddings like KannadaSBERT and TeluguSBERT significantly improve classification accuracy, showing a clear advantage in understanding and categorizing regional language text. This work addresses a critical gap in the field of NLP for regional Indian languages by providing a structured approach to classifying Kannada and Telugu news articles. It contributes to enhancing the organization and accessibility of information for Kannada - and Telugu-speaking communities, thereby advancing research in low- resource language processing.