Machine Intelligence and IOMT Based Heart Disease Prediction
Meenakshi Gupta, M. Amina Begum, Shaheen Ayyub, D. Baburao, Anita Soni, Vinod Patidar
Abstract
cardiac failure (HF) is a severe condition that can result from a number of different cardiac conditions. Patients with HF have wildly varying mortality rates, anywhere from 5% to 75%, depending on their prognosis. Assessing the overall mortality rate of HF patients is a crucial strategy for reducing premature deaths and improving health outcomes. The healthcare business and the early prediction of disease diagnosis have benefited greatly from IoMT and ML in present. Heart disease has always been leading killers. In light of this, this work proposes a multi-stage technique employing IoMT and ML for cardiac illness identification using both visual and numerical cues. In the first stage, features are extracted using transfer learning trained on a CNN. The best features are chosen in the second stage by employing one of three techniques: distributed-stochastic neighbour embedding (t-SNE), F-score/correlation-based feature-selection (CFS). The final step in cardiac patient diagnosis involves averaging the results from three different classifiers: the Gaussian-Bayes (GB), support-vector-machine (SVM), and random-forest (RF). The outcomes were tested using the aforementioned UCI datasets. Performance is shown to have increased in comparison to other approaches.