DZchatbot: A Medical Assistant Chatbot in the Algerian Arabic Dialect using Seq2Seq Model
Abdennour Boulesnane, Yaakoub Saidi, Oussama Kamel, Mohammed Mounir Bouhamed, Rostom Mennour
Abstract
In light of the global crisis like COVID-19, many people are afraid to leave the house and visit the doctor for fear of these epidemics. On the other side, the amazing development of artificial intelligence has led to chatbots' emergence and use in several fields. Therefore, in this paper, we propose to build an automated chatbot system that interacts with people in the Arabic Algerian dialect and helps patients ask general medical questions. To achieve this purpose, we propose three sequence-to-sequence models based on three Recurrent Neural Networks encoder-decoders: Long Short-Term Memory, Bidirectional Long Short-Term Memory, and Gated Recurrent Unit, to understand the user's request and provide the right useful answer. Experimentally, we have collected medical data of 2150 pairs. The results were very promising, and the proposed chatbot performed excellently in handling user questions.