Table 1.

The classification accuracies of LDA, Z-LDA, and EZ-LDA on simulation dataset.

Training size		20	30	40	50
LDA		78.7	80.4	81.9	83.4
Z-LDA		77.0	81.7	83.0	84.0
EZ-LDA (different thresholds)	10%	81.2∗	83.6∗	84.9∗	85.8∗
	20%	81.5∗	83.8∗	85.1∗	86.0∗
	30%	81.6∗	83.9∗	85.2∗	86.3∗
	40%	81.6∗	84.0∗	85.5∗	86.4∗
	50%	81.8∗	84.1∗	85.5∗	86.4∗
	60%	81.4∗	83.9∗	85.3∗	86.2∗
	70%	81.3∗	83.6∗	85.1∗	86.2∗
	80%	81.1∗	83.2∗	84.8∗	85.8∗
	90%	80.3∗	82.8∗	84.2∗	85.2∗

The second column denotes the threshold used in EZ-LDA.

∗ denotes the classification accuracy of EZ-LDA is significantly higher than that of Z-LDA (Mann-Whitney U test, p < 0.05).