A predictive prototype for the identification of diseases relied on the symptoms described by patients

Abstract

INTRODUCTION: A thorough and timely investigation of any health-related problem is essential for disease prevention and treatment. The normal way of diagnosis may not be sufficient in the event of a serious illness problem. OBJECTIVE: Creating a medical diagnosis prototype that uses many machine learning processes to forecast any illness relied on symptoms explained by patients can lead to an errorless diagnosis as compared to the traditional ways. METHODS: We created a disease prediction prototype using ML techniques such as random forest, CART, multinomial linear regression, and KNN. The data set utilized for processing contained over 132 illnesses. Diagnosis algorithm outcomes the ailment that the person may be suffering from relied on the symptoms provided by the patients. RESULTS: When compared to CART and random forest (accuracy is 97.72%, multinomial linear regression and KNN produced the best outcomes. The accuracy of the KNN prediction and multinomial linear regression techniques was 98.76%. CONCLUSION: The diagnostic prototype can function as a doctor in the early detection of an illness, ensuring that medical care can begin in an appropriate time and many lives can be secured.