Comparison Of Multinomial Naive Bayes Algorithm And Logistic Regression For Intent Classification In Chatbot

Abstract

Chatbot is software that communicates using natural language. chatbots such as machine conversation systems, Chatterbot, virtual agents, and dialogue systems. This software enables to simulate human conversations. In this research, the chatbot system that will be created must be able to understand the natural language of what is entered by the user, and the chatbot will answer according to what the user is expecting. The researcher proposes a classification method to identify intent rather than user input or called intent classification on the chatbot system; the researcher also wants to know the level of accuracy, precision, and recall on the evaluation results of both methods. The classification method applied in this research is the Naive Bayes method and compared with the Logistic Regression method to determine the class intention. The evaluation results show the level of accuracy precision and recall in the Logistic Regression model is higher than the Naive Bayes model.