Employability of the Machine Learning Algorithms in the Early Diagnosis of Various Diseases

Abstract

The most difficult task is accurately predicting disease. Environment and lifestyle factors contribute to a wide range of illnesses. As a result, it becomes a crucial task to predict disease earlier. On the other hand, the doctor finds it too difficult to predict symptoms accurately. Predicting the disease is important in using data mining to solve this issue. Medical science experiences significant annual data growth. Early patient care has benefited from accurate medical data analysis because of the growing amount of data in the medical field. Data mining uncovers hidden pattern information in a wide range of medical data by utilizing disease data. Based on the patient's symptoms, we proposed a general disease prediction. We use the machine learning algorithms K-Nearest Neighbor (KNN) and Convolutional Neural Network (CNN) for accurate disease prediction. A dataset of disease symptoms was required for disease prediction. A person's lifestyle and checkup information are considered for an accurate prediction in this general disease prediction. CNN has a higher general disease prediction accuracy of 84.5% than the KNN algorithm. Additionally, KNN's memory and time requirements are higher than CNN's. This system can provide the risk associated with the prevalent disease, which can be either a lower or higher risk of the prevalent disease after general disease prediction.