{"id":"arxiv/2607.16638","url":"https://arxiv.org/abs/2607.16638","tags":["math.ST","cs.LG","stat.ME"],"score":0,"title":"De-floored Principal Component Regression: When Rank Selection Alone Is Insufficient for Prediction","domain":"math.ST","source":"arxiv","authors":["Peng Zhao"],"reviews":0,"subject":"math.ST","summary":"Principal component regression (PCR) regularizes high-dimensional prediction by choosing a spectral cutoff, but rank selection cannot correct systematic inflation of the retained empirical eigenvalues. We study clean Gaussian random designs in which the aggregate covariance tail creates a nearly scalar sample-space floor comparable to the predictive head scale. De-floored principal component regression (dPCR) retains the cutoff and subtracts an estimated floor from the retained denominators. We prove an ordinary-PCR prediction-risk lower bound uniform over all ranks and a high-probability…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-07-18","reactions":{}}