An Automated Diagnostic System for Heart Disease Prediction Based on ${\chi^{2}}$ Statistical Model and Optimally Configured Deep Neural Network

Abstract

Different automated decision support systems based on artificial neural network (ANN) have been widely proposed for the detection of heart disease in previous studies. However, most of these techniques focus on the preprocessing of features only. In this paper, we focus on both, i.e., refinement of features and elimination of the problems posed by the predictive model, i.e., the problems of underfitting and overfitting. By avoiding the model from overfitting and underfitting, it can show good performance on both the datasets, i.e., training data and testing data. Inappropriate network configuration and irrelevant features often result in overfitting the training data. To eliminate irrelevant features, we propose to use χ 2 statistical model while the optimally configured deep neural network (DNN) is searched by using exhaustive search strategy. The strength of the proposed hybrid model named χ 2-DNN is evaluated by comparing its performance with conventional ANN and DNN models, another state of the art machine learning models and previously reported methods for heart disease prediction. The proposed model achieves the prediction accuracy of 93.33%. The obtained results are promising compared to the previously reported methods. The findings of the study suggest that the proposed diagnostic system can be used by physicians to accurately predict heart disease. INDEX TERMS Deep neural network, heart disease, hyperparameters optimization, overfitting, underfitting.

A hybrid cost-sensitive ensemble for heart disease prediction

“…Overall, the proposed model has following advantages compared with the stateof-the-art methods [48][49][50][51] :…”

“…Ethics approval and consent to participate Not applicable. 1 the result of the test, h = 1 indicates that the null hypothesis can be rejected at the 5% level 2 the probability of observing the given result by chance if the null hypothesis is true [56] Hybrid neural network 93 78.5 Shah et al [57] PPCA * + SVM 75 90.57 Marian and Filip [58] Fuzzy rule-based classification 84.70 92.90 Ali et al [48] Gaussian Naive Bayes classifier 87.80 97.95 Ali et al [49] Deep neural network 85.36 100 Ali et al [50] Hybrid SVM 82.92 100 Ali et al [51] Deep belief network 96.03 93.15 Arabasadi et al [59] Hybrid neural network-genetic algorithm 88 91 Mokeddem and Ahmed [41] Fuzzy…”

A hybrid cost-sensitive ensemble for heart disease prediction

“…Overall, the proposed model has following advantages compared with the stateof-the-art methods [50][51][52][53] :…”

“…The average performance on ten folds.The best result is bolded. [58] Hybrid neural network 93 78.5 Shah et al [59] PPCA 1 + SVM 75 90.57 Marian and Filip [60] Fuzzy rule-based classification 84.70 92.90 Ali et al [50] Gaussian Naive Bayes classifier 87.80 97.95 Ali et al [51] Deep neural network 85.36 100 Ali et al [52] Hybrid SVM 82.92 100 Ali et al [53] Deep belief network 96.03 93.15 Arabasadi et al [61] Hybrid neural network-genetic algorithm 88 91 Mokeddem and Ahmed [41] Fuzzy The values listed in the table represent the average performance on ten folds.…”

A hybrid cost-sensitive ensemble for heart disease prediction

“…Compared with the state-of-the-art methods [47][48][49][50], the proposed model has certain advantages: (1) It is easily understood by less experienced clinical physicians, which makes it easier to implement. (2) Considering different kinds of misclassification cost makes the proposed model closer to reality.…”