Server-based artificial intelligence robot analyzing COVID-19 symptoms

In the health sector, robots are used in various services by being strengthened by the Internet of Things and Artificial Intelligence-based studies. As of 2019, with the emergence of Covid-19, interest in the use of robots in infectious diseases and epidemic situations has increased. Within the scope of the fight against Covid-19; Various studies such as chat robot, disinfection robot, telemedicine robot, social distance robot were carried out. In this study, a server-based artificial intelligence-controlled CovidBot robot, which collects Covid-19 symptoms with internet of things technology, was made. The CovidBot system works with a robotic circuit, server and Naïve Bayes classification algorithm. CovidBot interacts with people and sends the data it obtains to the server by asking 13 questions to learn the symptoms of Covid-19. The classification process of the obtained data is done with the machine learning model created using Naïve Bayes, one of the supervised machine learning algorithms. For the training of the model, the training process of the machine learning model was carried out with three separate data sets. In the studies, more successful results were obtained with the machine learning model and the data set with more data than the training sets. This data set is classified with the trained model and divided into 4 different Covid groups as "Normal", "Mild", "Risky" and "Serious". In addition, if a suitable result cannot be obtained in the classification process, the relevant data is labeled as "Uncertain". These classification processes were evaluated with performance measures. In the classification study, an accuracy rate of 88% was obtained. Of course, the results obtained with CovidBot can be used in pre-diagnosis of Covid-19. Although it is advisory, it is capable of enabling people to get an idea without going into crowded test environments.