Towards a supervised classification of neocortical interneuron morphologies

BMC Bioinformatics

Table 2 Parameters for feature selection (KW and RF BVI), sampling (SMOTE) and cross-validation (CV)

Method	R Package	Parameters	Learner parameters
KW	stats [38]	α=0.05 adjust = FDR
RF BVI	randomForest [75]	bvi>0.01	\(T = 20000, m = \sqrt {n}\)
SMOTE	mlr [80]	k=5
CV	mlr [80]	r=10	k = 10, for ChC k=7

FDR stands for false discovery rate; r is the number of CV repetitions; k the number of folds. Learner parameters are the RF parameters used internally for RF BVI

ISSN: 1471-2105