Early Heart Disease Detection with Extreme Data Approach using Ensemble Learning Techniques

The heart is the most important organ in humans and has a big job to do. One of the most serious and persistent causes of death in the world today is heart disease. These are the results of poor prediction owing to ignorance. Cardiac attacks can be prevented if certain measures and safeguards are followed if these cardiac disorders can be predicted in advance. Manual prediction, however, may be extremely dangerous or perhaps impossible. Therefore, an earlier, quicker, and more effective technique for detecting cardiac disease is required. The proposed system involves utilizing ensemble learning strategies and multiple classifiers in its first stage to take care of the complicated problems with an extreme data to improve the efficacy and performance. The accuracy, precision, f1-score and recall of each classifier are calculated to estimate the efficiency of the classifier. With the classifier opted with best metrics opted, the presence of heart disease is detected. In the next stage the ECG signals of the patient in whom the heart disease is detected are analyzed by processing the signals into a series of convolutional layers. Precision, recall and f1-score are considered to understand categorization ability of the model.