A COMPREHENSIVE ANALYTICAL METHOD FOR PREDICTING DISEASES THROUGH MACHINE LEARNING APPROACHES

In the area of data mining, supervised machine learning techniques have been a prevalent approach. Recently, a possible use case for these techniques has emerged:disease prediction utilising health data. Early identification of infectious illnesses is thus crucial, and several researchers have created models to identify them early. It requires a lot of conventional clinical studies, which might make predicting the illness more difficult. A thorough analysis of several illness prediction techniques is provided in this work. The present research offers a prediction model that makes use of many information combinations and well-known categorisation techniques. In order to improve the prediction potential of the model, techniques like the Under-Sampling Clustering Over sampling Method (USCOM) address the issue of data imbalance. Adaptive Elephant Herd Optimisation Method (AEHOM) training is used to a multilayeredDeep Convolutional Neural Network (MLDCNN) for the classification job. According to its comprehensive analysis, the proposed Machine Learning-Based Heart Disease Prediction Method (ML-HDPM) has excellent results across many significant assessment variables.The ML-HDPM model performs very well during training, achieving 94.8% precision and 95.5% accuracy. With an accuracy rate of 96.2%, the system's sensitivities (Recall) are very accurate, and its F-score of 91.5% highlights its well-balanced performance. An excellent 89.7% specificity is reported by ML-HDPM, which is significant. The findings demonstrate how ML-HDPM may transform the prediction of heart disease and assist medical practitioners in making precise diagnoses, significantly impacting patient care results.