Comparison of Rule-based Chat Bots with Different Machine Learning Models

A chat bot is an assistant software which is capable of performing conversations with human users in textual or speech format. They can input, understand the context behind it and generate an appropriate response as per the query asked to them. The primary objective of this study is to evaluate and compare the performance and effectiveness of rule-based chat bot systems utilizing different machine learning approaches. The research methodology involves the implementation of multiple chat bot systems based on rule-based architectures, where each system is powered by a distinct machine learning model. These models encompass a range of techniques such as NLP, traditional machine learning algorithms (e.g., Decision trees, Naive Bayes, etc.), and deep learning models (e.g. sequential ANN). Furthermore, to enhance transparency and interpretability of the chat bot systems, the paper explores the integration of instance explanations using LIME.