cov

→Tensor

cov(input: Tensor, correction: int = 1)

Covariance matrix of input.

inputTensor

Shape (N,) (single-variable degenerate case) or (N, M) — N variables observed at M time points.

correctionint= 1

Bessel correction; the divisor is max(1, M - correction). Default 1 (unbiased sample covariance); pass 0 for the biased population covariance.

Tensor

Shape (N, N). Entry $(i, j)$ = $\frac{1}{M - \text{correction}}\sum_k (x_{i,k} - \bar x_i)(x_{j,k} - \bar x_j)$ .

→Tensor

cov(input: Tensor, correction: int = 1)

Covariance matrix of input.

inputTensor

Shape (N,) (single-variable degenerate case) or (N, M) — N variables observed at M time points.

correctionint= 1

Bessel correction; the divisor is max(1, M - correction). Default 1 (unbiased sample covariance); pass 0 for the biased population covariance.

Tensor

Shape (N, N). Entry $(i, j)$ = $\frac{1}{M - \text{correction}}\sum_k (x_{i,k} - \bar x_i)(x_{j,k} - \bar x_j)$ .