Fall risk classification with posturographic parameters in community-dwelling older adults: a machine learning and explainable artificial intelligence approach

Table 4 Output using Criteria I, at risk of fall according to fall history in the past 1 year, as a classification criterion

Feature selection	Model	Accuracy	Recall	Specificity	AUC
ABC	Balanced Bagging	0.69	0.73	0.68	0.70
	Complement NB	0.69	0.67	0.70	0.71
	Easy Ensemble	0.69	0.67	0.70	0.71
HHO	Balanced Bagging	0.60	0.47	0.64	0.61
	Complement NB	0.60	0.40	0.66	0.63
	Easy Ensemble	0.60	0.40	0.66	0.63
SMA	Balanced Bagging	0.58	0.40	0.64	0.49
	Complement NB	0.62	0.53	0.64	0.64
	Easy Ensemble	0.60	0.60	0.60	0.66
None	Balanced Bagging	0.71	0.60	0.74	0.72
	Complement NB	0.71	0.60	0.74	0.72
	Easy Ensemble	0.71	0.60	0.74	0.72
Mean	Balanced Bagging	0.65	0.55	0.68	0.63
	Complement NB	0.66	0.55	0.69	0.68
	Easy Ensemble	0.65	0.57	0.68	0.68

AUC: area under the curve; ABC: Artificial Bee Colony; HHO: Harris Hawk Optimization; SMA: Slime Mould Algorithm

ISSN: 1743-0003