Class SingularValueDecomposition

constructor

• new SingularValueDecomposition(value, options?): SingularValueDecomposition

• Parameters

  • value: MaybeMatrix

  • Optional options: ISVDOptions

  Returns SingularValueDecomposition

Readonly condition

Readonly diagonal

Readonly diagonalMatrix

Readonly leftSingularVectors

Readonly norm2

Readonly rank

Readonly rightSingularVectors

Readonly threshold

inverse

• inverse(): Matrix

• Get the inverse of the matrix. We compute the inverse of a matrix using SVD when this matrix is singular or ill-conditioned. Example : var svd = SingularValueDecomposition(A); var inverseA = svd.inverse();

  Returns Matrix

  • The approximation of the inverse of the matrix.

solve

• solve(value): Matrix

• Solve a problem of least square (Ax=b) by using the SVD. Useful when A is singular. When A is not singular, it would be better to use qr.solve(value). Example : We search to approximate x, with A matrix shape m*n, x vector size n, b vector size m (m > n). We will use : var svd = SingularValueDecomposition(A); var x = svd.solve(b);

  Parameters

  • value: Matrix

    Matrix 1D which is the vector b (in the equation Ax = b).

  Returns Matrix

  • The vector x.

solveForDiagonal

• solveForDiagonal(value): Matrix

• Parameters

  • value: ArrayLike<number>

  Returns Matrix