Table 1.

Overview of the filter methods: Name of the filter method (filter), short description of filter (description), information if filter is multivariate (multivariate), information if filter uses the survival outcome or a transformed target variable (target), information if filter requires categorization of numeric features (features category), and information about R package from which the implementation is taken

Filter	Description	Multivariate	Target	Features category	Implementation
variance	Feature variance	No		No	mlr3filters [60]
correlation	Pearson correlation	No		No	mlr3filters [60]
cox.score	Score test	No		No	survival [61]
carss	Correlation-adjusted regression survival scores	No		No	carSurv [62]
permutation	Random forest permutation importance	Yes		No	ranger [63] with default hyper parameter settings
impurity	Random forest impurity importance	Yes		No	ranger [63] with default hyper parameter settings
boosting	Boosting importance	Yes		No	xgboost [64] with 2 000 boosting iterations, step size 0.05 and maximum tree depth 10
mim	Mutual information	No		Yes	praznik [55]
mrmr	Mutual information	Yes		Yes	praznik [55]
jmi	Mutual information	Yes		Yes	praznik [55]
jmim	Mutual information	Yes		Yes	praznik [55]
disr	Mutual information	Yes		Yes	praznik [55]
njmim	Mutual information	Yes		Yes	praznik [55]
cmim	Mutual information	Yes		Yes	praznik [55]