A Clinical Decision Support System for Heart Disease Prediction Using Deep Learning

Unfortunately, heart disease is currently the primary cause of mortality worldwide and its incidence is increasing. Detecting heart disease in its initial stages before a cardiac event takes place poses challenges. Huge amount of heart disease data is available in the health care sector such as in clinics, hospitals etc. However, this data is not intelligently handled to identify the hidden patterns. Machine learning techniques help in turning this medical data into useful knowledge. Machine learning is used to design such decision support systems (DSS) that can learn and improve from their past experiences. Recently, deep learning has gained the interest of industry and academics. The fundamental objective of this research activity is the precise diagnosis of heart illness. The suggested approach makes use of a Keras-based deep learning model to compute results with a dense neural network. The proposed model undergoes testing with various configurations of hidden layers in the dense neural network, ranging from 3 layers to 9 layers. Each hidden layer employs 100 neurons and utilizes the Relu activation function. To carry out the analysis, several heart disease datasets are utilized as benchmarks. The assessment encompasses both individual and ensemble models, and is performed on all heart disease datasets. Furthermore, using important measures like sensitivity, specificity, accuracy, and f-measure, the dense neural network is assessed across all datasets. The performance of different layer combinations varies across datasets due to varying attribute categories. Through extensive experimentation, the results of the proposed framework are analyzed. The study's conclusions show that, when applied to all heart disease datasets, the deep learning model suggested in this research paper achieves superior accuracy, sensitivity, and specificity compared to individual models and alternative ensemble approaches.