Design and Implementation of an Assamese Language Chatbot Using Neural Networks

This research paper proposes a design for an AI chatbot that can understand Assamese language using a feedforward neural network. We have developed an Assamese language dataset and trained the model on it. We have used the bag of words method for feature extraction of the text. We have evaluated the accuracy and performance of the model using metrics like precision, recall and F1 score. The model has achieved an accuracy of 89.11%, which is higher than SVM and Naive Bayes models. We have also presented the future plan of this research. The novelty of this research is that we have developed a deep learning FFNN model for an AI Assamese chatbot along with an Assamese dataset, which has not been done before. The significance of this research is that it will help to explore the AI Assamese chatbot domain and inspire new ideas for researchers, as well as help the Assamese society to use their language in digital platforms.