AZOrange - High performance open source machine learning for QSAR modeling in a graphical programming environment

Table 1 Optimized model hyper-parameters

Algorithm	Parameter	Range
RF	The number of active attributes
SVM	C	2^-5 to 2¹⁵
	γ	2^-15 to 2³
ANN	The number of hidden neurons (one-layer)
CvBoost	Maximum branching depth	1 to 20
	The number of trees	1 to 1000
PLS	The number of components	1 to nAttr-1

Model hyper-parameters being optimized by default and their corresponding ranges.
¹nAttr is the number of attributes in the data set
²nEx is the number of examples in the data set

ISSN: 1758-2946