Infectious Disease Prediction, Testing Suggestionfor Better Operational Health Careusing Machine Learning Algorithms

Abstract

People are suffering from different types of infectious diseases such as malaria, dengue, corona, chikungunya,etc. “Pathogenic microorganisms such as bacteria, viruses, parasites, and fungi are the primary cause of infectious diseases”.For all these infectious diseases the main symptom is fever, and people suffer from other symptoms such as Dehydration, Chills, Joint Pain, Muscle Pain, Head Ache, Joint Swelling, Rash, Vomiting, Cough, Shortness of Breathing, Chest Pain, Runny and Watery Nose, Loss of Appetite, etc. Medical practitioners are suggesting various tests to confirm the type of disease. If people are aware of which type of test is necessary to confirm the disease typedepending on the symptoms or signs of the patient and domain knowledge, then the testing expenditures can be reduced for the patients. So, in this paper,a hybrid model thatcombines classification and association analysis was developed that predicts the type of infection based on symptoms of patients and also suggests the test to be performed to confirm the disease. Naïve Bayes, SMO, and Apriori algorithms were implemented in the hybrid model. All models performed equally well with 92% of prediction accuracy. The model finds various associations among symptoms, disease, and tests performed which can be used for the testing suggestion. Further, the model can be extended to implement it in real-time by developing a web application.