Function qrSolve

qrSolve(
    a: MatrixInput | QR,
    b?: VectorOrMatrix,
    options?: { tol?: number },
): number[] | number[][]
Solves a x = b in the least-squares sense using a QR decomposition. Equivalent to R's qr.solve().

Throws if the rank is less than min(nrow, ncol). When b is omitted, returns the inverse of a (square matrices only). For a wide matrix, coefficients of unused columns are set to 0 (as in R).

A LAPACK decomposition (qr(a, { lapack: true })) has no rank, so it throws instead if some |R[k][k]| is at most tol * |R[0][0]| (tol from options, default 1e-7), which R's qr.solve() does not detect.
Parameters
- a: MatrixInput | QR
  A matrix (array of rows) or a QR decomposition from qr.
- Optionalb: VectorOrMatrix
  The right-hand side, a vector or a matrix. Default: the identity matrix.
- options: { tol?: number } = {}
  tol: the tolerance passed to qr when a is a matrix. Default: 1e-7.
Returns number[] | number[][]
A vector if b is a vector, otherwise an array of rows.
Example
```
qrSolve([[2, 1], [1, 3]], [3, 5]); // [0.8, 1.4]
qrSolve([[1, 1], [1, 2], [1, 3]], [1, 2, 4]); // [-0.6667, 1.5]
```
- Defined in linalg/qr.ts:793