Skip to main content

Table 4 Survival prediction results on all clinical features – mean of 100 executions

From: Machine learning can predict survival of patients with heart failure from serum creatinine and ejection fraction alone

MethodMCCF1 scoreAccuracyTP rateTN ratePR AUCROC AUC
Random forestsblue+0.384*0.547blue0.740*0.4910.8640.657blue0.800*
Decision tree+0.376blue0.554*0.737blue0.532*0.8310.5060.681
Gradient boosting+0.3670.5270.7380.4770.8600.5940.754
Linear regression+0.3320.4750.7300.3940.8920.4950.643
One rule+0.3190.4650.7290.3830.8920.4820.637
Artificial neural network+0.2620.4830.6800.4280.815blue0.750*0.559
Naïve bayes+0.2240.3640.6960.2790.8980.4370.589
SVM radial+0.1590.1820.6900.1220.9670.5870.749
SVM linear+0.1070.1150.6840.072blue0.981*0.5940.754
k-nearest neighbors-0.0250.1480.6240.1210.8660.3230.493
  1. MCC: Matthews correlation coefficient. TP rate: true positive rate (sensitivity, recall). TN rate: true negative rate (specificify). Confusion matrix threshold for MCC, F1 score, accuracy, TP rate, TN rate: τ=0.5. PR: precision-recall curve. ROC: receiver operating characteristic curve. AUC: area under the curve. MCC: worst value = –1 and best value = +1. F1 score, accuracy, TP rate, TN rate, PR AUC, ROC AUC: worst value = 0 and best value = 1. MCC, F1 score, accuracy, TP rate, TN rate, PR AUC, ROC AUC formulas: Additional file 1 (“Binary statistical rates” section). Gradient boosting: eXtreme Gradient Boosting (XGBoost). SVM radial: Support Vector Machine with radial Gaussian kernel. SVM linear: Support Vector Machine with linear kernel. Our hyper-parameter grid search optimization for k-Nearest Neighbors selected k=3 on most of the times (10 runs out of 100). Our hyper-parameter grid search optimization for the Support Vector Machine with radial Gaussian kernel selected C=10 on most of the times (56 runs out of 100). Our hyper-parameter grid search optimization for the Support Vector Machine with linear kernel selected C=0.1 on most of the times (50 runs out of 100). Our hyper-parameter grid search optimization for the Artificial Neural Network selected 1 hidden layer and 100 hidden units on most of the times (74 runs out of 100). We report bluein blue and with the top performer results for each score.