math.STJun 9, 2026

Bidirectional Random Projections

Authors: Chao Lan, Luyuan Yang

Organizations: a School of Computer Science, University of Oklahoma, U.S.

Abstract

This paper analyzes bidirectional random projections for ordinary least squares (OLS) regression under the fixed design setting. Let (X,Y)∈Rn×p×Rn(X,Y) \in \mathbb{R}^{n \times p} \times \mathbb{R}^n be a sample and R∈Rn1×n,W∈Rp×p1R \in \mathbb{R}^{n_1 \times n}, W \in \mathbb{R}^{p \times p_1} be two properly distributed random projections. We develop an expected excess loss bound for the OLS estimator built on (WXR,WY)(WXR, WY). Compared to an established bound for OLS estimator built on (XR,Y)(XR, Y), the gap is approximately O(p1+C1p1)O\left( p_1 + C \frac{1}{p_1} \right), where CC scales with n1/nn_1/n and can be negative for small n1/nn_1/n. Its implications are confirmed by numerical results on real-world data.

Explore similar work

CardsList
  1. Distributed Sketching on Data Partitions for OLS Regression

    Jul 8, 2026Luyuan Yang, Brayden Garner, Shayan Shafaei +1Least SquaresPartitioning