A Decision-Making System for Clinical Data

Heart disease is a leading global killer, yet it may often be prevented if diagnosed and treated early. When diagnosing a patient for heart disease, a clinical decision support system (CDSS) could be useful. Clinical decision-making supported by deep learning (DL) has recently been used in the medical industry. According to existing research, decision-making systems based on DL techniques have been proved to be effective in the prediction of heart illness in individuals. Outlier identification and data normalisation methods have not been used in any of the existing investigations, especially when it comes to heart disease datasets. This study proposes a novel weighted fuzzy-based recurrent neural network (WFRNN) with a genetic algorithm (GA) for a CDSS. A boosted K-means clustering (BKC) is employed to identify and remove outliers. Training data propagation can be evenly distributed by using a synthetic minority oversampling (SMOTE-Out) technique. Heart disease can be predicted using the proposed WFRNN, which incorporates the GA. The statlog dataset was employed in this study. The proposed model scored better than those of existing models and earlier studies.