Bidirectional Long Short-Term Memory Model for Intent Classification in Customer Service Chatbot

The demand for responsive and efficient customer service is a crucial aspect of enhancing customer satisfaction, particularly in Indonesian government offices abroad. To address this challenge is implementing a chatbot system based on Bidirectional Long Short-Term Memory. This model can understand conversational contexts more comprehensively, enabling it to generate relevant and timely responses. This study aims to optimize chatbot performance in enhancing customer experience by implementing the Bi LSTM algorithm to handle intent classification of customer input data. Experimental results demonstrate that this model successfully improves evaluation metrics, achieving an accuracy of 84.64%, precision of 85%, recall of 85%, and an F1-score of 85%.