Evaluating the Effectiveness of GPT-3.5 and BERT in Classifying Bangla News Articles

This paper evaluates the effectiveness of GPT-3.5 and BERT models in classifying Bangla news articles. The study involves collecting a dataset of Bangla news articles, preprocessing the data, using the models to generate predictions, and comparing their performance. The performance of the models is assessed using metrics such as accuracy, precision, recall, and F1-score. The results indicate that the BERT model outperforms GPT-3.5, achieving higher accuracy and better precision and recall scores. The findings highlight the importance of tailored natural language processing (NLP) solutions for low-resource languages like Bangla and suggest that transformer-based models can effectively capture the nuances of the Bangla language when adapted appropriately. This research contributes to the broader field of NLP for under-resourced languages.