. 2026 May 27;16:16572. doi: 10.1038/s41598-026-53824-6

Table 3.

Confusion matrices from each logistic regression model.

Model	Observed			Accuracy
	Prediction	Male	Female	Male	Female	Overall
Model 1: Verbal-Episodic	Male	582	395	0.51	0.70	0.61
Model 1: Verbal-Episodic	Female	568	929	0.51	0.70	0.61
Model 2: Verbal-Episodic + Spatial	Male	789	314	0.69	0.76	0.73
Model 2: Verbal-Episodic + Spatial	Female	361	1010	0.69	0.76	0.73
Model 3: Verbal-Episodic + Spatial + Personality	Male	847	277	0.74	0.79	0.77
Model 3: Verbal-Episodic + Spatial + Personality	Female	303	1047	0.74	0.79	0.77
Model 4: Verbal-Episodic + Spatial + Personality + Interests	Male	886	234	0.77	0.82	0.80
	Female	264	1090	0.77	0.82	0.80

The cells with predictions list predicted values (where values above 0.5 are classified as ones and below 0.5 are classified as zeroes) and compare these with observed values. The diagonal represents correct predictions and the off-diagonal incorrect predictions.