Supporting Risk-Aware Use of Online Translation Tools in Delivering Mental Healthcare Services among Spanish-Speaking Populations

Table 5

Performance of RVM classifiers with all (dispersion + USAS + LIWC + structural) features.

RVM	Training data		Testing data
	AUC mean	SD	AUC	Accuracy	Sensitivity	Specificity
ALL 251	0.642	0.038	0.647	0.613	0.518	0.706
ALL 251 with min-max	0.626	0.041	0.697	0.643	0.590	0.694
ALL 251 with L2	0.670	0.045	0.653	0.619	0.639	0.600
ALL 251 with Z-score	0.633	0.085	0.680	0.625	0.590	0.659
ALL33 (automatically optimised)	0.658	0.014	0.670	0.649	0.554	0.741
ALL33 with min-max	0.678	0.034	0.718	0.696	0.651	0.741
ALL33 with L2	0.710	0.015	0.670	0.637	0.651	0.624
ALL33 with Z-score	0.672	0.036	0.682	0.643	0.627	0.659