MoDeSuS: A Machine Learning Tool for Selection of Molecular Descriptors in QSAR Studies Applied to Molecular Informatics

Table 3

Statistical results for fs-randomization and y-randomization and performance of the model inferred from descriptors of Subset_1C (). For each case, for the percentage of cases correctly classified (%CC) and the Matthews Correlation Coefficient (MCC), three metrics are reported: mode, variance, and percentile₉₉.

fs – randomization
	Mode	Variance	Perc(99)
%CC	79.85	8.81	86.71	87
MCC	0.53	0.01	0.66	0.67
y – randomization
	Mode	Variance	Perc(99)
%CC	63.73	4.10	68.05	87
MCC	0.03	0.001	0.09	0.67