Data for IQC Paper

<p dir="ltr">All data generated and used in the manuscript "<i>IQC</i>: A Novel Criterion for Assessing Feature Selection Stability in High-Dimensional Analyses." General information on each file is described below:</p><p dir="ltr"><b><u>Simulated Data</u></b></p><ul><li>"IQC_Model_Iterations_[...].csv": Files containing all 100 model iterations used to compute IQC for microarray gene expression case study. File listed as "Unfiltered" contains model iterations from run without filtering genes with low IQR.</li><li>"IQC_Results_All_Simulations.csv": Contains IQC values for all simulations, across every combination of signal-to-noise ratio and number of predictors.</li><li>"Microarray_Submodel_Alphas_Filtered_Full.csv": Contains values for hyperparameter α for both filtered and unfiltered microarray IQC models.</li><li>"S1_Fig_Data.csv": Contains data necessary to reproduce S1 Fig.</li><li>"Simulated_Data_[...]_FTN_[...]_Feats.csv": Full simulated data sets for each configuration of signal-to-noise and total number of predictors used in manuscript.</li><li>"TNR_TPR_Rates_[...].csv": True negative and true positive rates for each simulation from IQC manuscript.

Used to generate Fig 2.</li></ul><p dir="ltr"><b><u>Microarray Data</u></b></p><ul><li>"golub.csv": Microarray gene expression data set used to classify leukemia subtype. Data originally gathered and published in Golub et al. (1999), and is publicly available through Bioconductor at the following DOI: 10.18129/B9.bioc.golubEsets.</li></ul><p></p>