Optimization of Classification Strategies of Acetowhite Temporal Patterns towards Improving Diagnostic Performance of Colposcopy

Table 3

Performance of automatic classification methods using acetowhite temporal patterns on a dataset of 200 cases.

Method	Model	Classified instances (%)		TP rate	FP rate	Precision	Recall	F-Measure	MCC	ROC area	PRC area
		Correctly	Incorrectly
IBk	Standardized	70	30	0.700	0.309	0.700	0.700	0.699	0.395	0.721	0.683
	Adjusted	69	31	0.690	0.319	0.689	0.690	0.689	0.375	0.717	0.679
	Parameters	64	36	0.635	0.375	0.634	0.635	0.633	0.263	0.632	0.599
	PLA	70	30	0.700	0.313	0.701	0.700	0.697	0.395	0.732	0.713
	PSA	62	38	0.620	0.437	0.778	0.620	0.539	0.327	0.610	0.608
Naïve Bayes	Standardized	69	31	0.690	0.329	0.695	0.690	0.684	0.377	0.713	0.681
	Adjusted	69	31	0.685	0.333	0.689	0.685	0.679	0.366	0.708	0.673
	Parameters	53	47	0.525	0.459	0.537	0.525	0.520	0.067	0.540	0.543
	PLA	65	35	0.645	0.385	0.658	0.645	0.627	0.289	0.697	0.669
	PSA	61	39	0.605	0.409	0.603	0.605	0.601	0.200	0.624	0.617
C4.5	Standardized	68	32	0.675	0.330	0.674	0.675	0.675	0.346	0.627	0.592
	Adjusted	64	36	0.640	0.361	0.641	0.640	0.640	0.279	0.618	0.582
	Parameters	55	45	0.545	0.471	0.541	0.545	0.539	0.076	0.526	0.520
	PLA	65	35	0.645	0.361	0.644	0.645	0.645	0.285	0.652	0.611
	PSA	64	36	0.635	0.379	0.634	0.635	0.631	0.262	0.643	0.611