Abstract:
Covid-19 has touched almost all of the world’s population lives and still a lot of people are being
affected throughout the world including our country Ethiopia. Chatbot systems are becoming an
emerging technology and are supporting individuals and experts in health sectors. The objective
of this study is to develop nutrition assistance Chatbot system for Covid-19 patients that can
recommend the type and quantity of foods the patients should take by collecting different
parameters from the patients as an input. In this research the researcher motivated by that no
research work has been conducted using this method and conducted research paper aims to
address the challenges posed by the COVID-19 pandemic. The researcher Focused on the
nutritional aspects and the development of a Chatbot system to assist patients, medical
professionals, and researchers. Since there is no nutritional recommendation found on patient’s
records the researcher collected the data from medical doctors, nurses and nutritionists from
Jimma University Health center and World Health Organizations nutritional recommendations
were collected to prepare a total of 933 dataset which is organized as a JavaScript Object
Notation file. The collected data’s were nutritional recommendations for Covid-19 patients.
These data’s were collected by interviewing the medical doctors and nutritional
recommendations from world health organization. The model training process utilized the
Convolutional Neural Network architecture; with hyper parameter tuning achieved a training
accuracy of 99.55%. Overall the developed system is effective in for recommending nutritional
intakes for the COVID-19 patients. The feedback received from domain experts further validated
the system's performance, indicating its potential utility in providing accurate and relevant
nutritional recommendations. The researcher is now working on post editing to enhance the
performance of the Chatbot system to give better service for the users.
