Using Healthcare Resources Wisely: A Predictive Support System Regarding the Severity of Patient Falls

Table 5

Model performance assessments.

Dataset	Learner	Accuracy (SD)	F1 (SD)	Precision (SD)	Recall (SD)
Training	Multinomial logistic regression (MLR)	0.442 (0.028)	0.442 (0.028)	0.443 (0.029)	0.443 (0.028)
	Naïve Bayes (NB)	0.461 (0.026)	0.448 (0.026)	0.460 (0.028)	0.472 (0.024)
	Random forest (RF)	0.783 (0.008)	0.784 (0.007)	0.785 (0.007)	0.788 (0.008)
	Support vector machine (SVM)	0.771 (0.008)	0.771 (0.008)	0.774 (0.008)	0.776 (0.009)
	eXtreme gradient boosting (XGBoost)	0.778 (0.006)	0.779 (0.005)	0.781 (0.005)	0.784 (0.005)
	Deep learning (DL)	0.721 (0.016)	0.720 (0.017)	0.735 (0.013)	0.725 (0.019)
	Stacking (RF + SVM + XGBoost + DL)	0.756 (0.014)	0.754 (0.014)	0.760 (0.019)	0.763 (0.015)
Test	Multinomial logistic regression	0.426	0.397	0.402	0.416
	Naïve Bayes	0.426	0.426	0.444	0.500
	Random forest	0.844	0.850	0.839	0.875
	Support vector machine	0.823	0.828	0.817	0.851
	eXtreme gradient boosting	0.835	0.843	0.831	0.866
	Deep learning	0.751	0.743	0.725	0.773
	Stacking (RF + SVM + XGBoost + DL)	0.781	0.775	0.758	0.799

Note. SD denotes standard deviation.