Given a least square problem Ax≃bA \textbf{x} \simeq \textbf{b}Ax≃b, we obtain
This can be viewed as a new linear system Cx^=dC \hat{\textbf{x}} = \textbf{d}Cx^=d, where a square matrix C=ATA∈Rn×nC = A^T A \in \mathbb{R}^{n \times n}C=ATA∈Rn×n, and d=ATb∈Rn\textbf{d} = A^T \textbf{b} \in \mathbb{R}^nd=ATb∈Rn