Intelligent Identification of Respiratory Diseases: Covid-19 and Similar Virus Cases.

The World Health Organization (WHO) has declared an end to the emergency phase of the COVID-19 pandemic. However, it is still negatively impacting countries like Ethiopia. Thus, exploring the identification and treatment mechanisms for COVID-19 from similar disease has still paramount importance. COVID-19 testing is still expensive, not easily available and time consuming, particularly in the conflict hit Tigray region. In this study, our main goal is to explore, design, implement and evaluate intelligent software that identifies COVID-19 from similar diseases, namely, Common Cold, Measles, Flu, COVID-19, and Chicken Pox. Furthermore, we point out the benefits of implementing such software to replace shortage of manpower given the circumstances in Tigray region. The software, hereafter named CKBS (COVID-19 Knowledge-based System), has been evaluated to ensure whether it is accurate and usable. In this study, 25 patients participated for each of the five different diseases to test the effectiveness of the developed knowledge base. The dataset was split in such a way that 70% is used for training while the remaining 30% for testing. Accordingly, we found out that the resulting software can successfully identify the five diseases at 98% accuracy and getting 96% community acceptance rate. This implies that our use of the tree model in the development was effective and resulting software has been accepted to be used by the local community to tackle the COVID-19 and related challenges that Ethiopia (particularly Tigray) is still facing.