TABLE 3.

Commonly used feature selection methods by the studies included in this review

Feature selection method	Mechanism
Component Analysis	Variance via sorted eigenvalues
Clustering	Choosing representative features among correlated groups
ICC/CCC	Measures feature reproducibility with correlation
LASSO	Regression analysis with L1 regularization
mRMR	Maximize F‐statistic and minimize correlation with defined feature limit
(Non)Parametric Statistics	Analysis of variance or means between two or more datasets
Pearson/Spearman Correlation	Determines highly correlated features prior to feature selection
Rank Sum	Two‐sided median analysis
Regression	Statistical relationship between dependent and independent variables
Relief	Scoring based on the nearest neighbor feature value differences
Random Forest	Calculate importance according to pureness of leaves

Abbreviations: CCC, concordance correlation coefficient; ICC, intraclass correlation coefficient; LASSO, least absolute shrinkage and selection operator.