Volume 18, No. 5, 2021
Developing a Machine Learning-Based Framework for Disease Prediction
Kumud Pant , Dibyahash Bordoloi , Bhasker Pant
Abstract
Due to environmental conditions and contemporary lifestyles, individuals encounter a variety of ailments presently. Predicting diseases at an early stage thus becomes an essential endeavor. However, the precise prognosis from symptoms becomes increasingly challenging for doctors. Correctly predicting sickness is the most difficult job. To solve this issue, data mining plays a crucial role in predicting the illness. The health sector generates an increasing quantity of data each year. Due to the rise in the quantity of information available in the healthcare field, initial patient treatment has benefited from the thorough assessment of medical data. Massive amounts of patient records are mined for hidden patterns with the use of disease-specific data. We presented a broad illness prediction system depending on the patient's health condition. We employ the K-Nearest Neighbor (KNN) and Convolutional neural network (CNN) machine learning algorithms for effective illness prediction. Knowledge of illness symptoms is essential for prediction and diagnosis. In this generalized illness prediction, the individual's lifestyle and test results are considered for an effective prediction outcome. CNN's general illness accuracy rate is 89.3%, that is higher than the KNN approach. Additionally, KNN is more demanding in terms of time and storage space than CNN. Following general illness forecasting, this method can provide the risk linked to the overall disease, whether that risk is low or high.
Pages: 3132-3139
Keywords: Machine Learning approaches, Disease Forecasting, Healthcare sector, KNN, CNN.