Utilizing AI for viral infection diagnosis: a case study in Tigray, Ethiopia

In recent years, the emergence of viral infections has posed significant challenges to healthcare systems worldwide. Timely and accurate detection of viral infections is essential for effective patient care and outbreak management. Although the World Health Organization has declared an end to the emergency phase of the Covid-19 pandemic, it is still impacting countries like Ethiopia, particularly in the war-ridden Tigray region. Today, viral infection testing in Tigray is still expensive, less accessible and time consuming. In this study, we introduce a novel approach for intelligent detection of viral infections, particularly Covid-19 from similar other diseases. A rule-based system, called CSDIS, was developed using tree to identify symptoms of the viruses and analyze the behavior of each diseases. Accuracy and usability of the system was evaluated by employing a dataset of 100 patient cases. We use 70% of the dataset for training, and the remaining 30% for testing the model. Additionally, standard RT-PCR tests are employed to confirm the model prediction results. Accordingly, the developed model proved successful in accurately identifying each class of disease with 98% accuracy, and 96% acceptance rating from the local community. This suggests that the system can empower individual healthcare professionals and minimize the shortage of manpower and resources. Overall, this research not only enhances the efficiency and precision of viral infection detection but also has the capacity to transform how healthcare professionals, researchers, and policymakers respond to viral outbreaks.