Comparative Analysis of Intent Classification in Indonesian Chatbots Using BERT and RoBERTa Models

Achatbot is a software application to designed handle user inputs and generate appropriate replies based on those inputs, which are then communicated back to the user. To ensure effective response from users, need to correctly interpret the intent chatbot. This involves identifying the underlying meaning of the text provided by the user, allowing the chatbot to deliver relevant responses. This paper compares two state-of-the-art transformer-based models-BERT (Bidirectional Encoder Representations from Transformers) and RoBERTa (Robustly Optimized BERT Pretraining Approach)-for the task of intent classification in chatbot systems. Various performance metrics, including accuracy, F1-score, precision, and recall, were analyzed to determine which model performs more effectively under different conditions. Performance metrics like accuracy and F1-score were compared to assess model BERT and RoBERTa performs better in a University Chatbot Dataset in Indonesian language. The BERT model achieved an accuracy of 0.89, outperforming the RoBERTa model, which achieved 0.84.