Cardiovascular Diseases Prediction Based on Dense-DNN and Feature Selection Techniques

Cardiovascular Diseases (CVDs) are a group of disorders affecting the heart and blood vessels. They have been considered in recent years as one of the main causes of death in the world. Patients with heart disease do not feel sick until the very last stage of the disease and most heart patients die before receiving any treatment. Machine Learning and Deep Learning techniques play an important role in early prediction of heart disease, to improve the quality of healthcare and help individuals to avoid earlier health complications as coronary artery infection and decreased function of blood vessels . Nowadays, the field of health care produces a large amount of data. The need for efficient techniques for processing this data has become necessary. In this paper, a model for cardiovascular disease prediction based on Dense Deep Neural Networks (Dense-DNN) is developed and attributes selection is performed via a Genetic Algorithm (GA). The GA is used to identify the best subset of attributes from the entire features in the dataset, to improve the performances and reduce the training time of the classification model. Our prediction model is compared to several traditional Machine Learning techniques. The performances of our system have been evaluated based on six parameters: (1) accuracy, (2) sensitivity, (3) specificity, (4) F-measure, (5) RMSE, and (6) MAE. Experimental results show that our proposed model outperforms state-of-the-art methods in terms of performance evaluation metrics. The achieved accuracy of the proposed model is 91.7% without using feature selection and 95% with the use of feature selection.