Fig. 1From: A generalized covariate-adjusted top-scoring pair algorithm with applications to diabetic kidney disease stage classification in the Chronic Renal Insufficiency Cohort (CRIC) StudyComparison of residualizing feature pairs from the first simulation study. Left column: Scatter plots of generated feature pairs from our first simulation study (N = 200) conditional on our single “clinical” covariate, \((X_{1} ,{ }X_{2} )\), and independent of our single “clinical” covariate, \((X_{3} ,{ }X_{4} )\). Right column: Scatter plots of the residualized feature pairs. The two evenly split classes are represented as red and blue and TSP’s decision boundary is overlayed on the plotsBack to article page