Procs
proc covariance_matrix[T: SomeFloat](x, y: Tensor[T]): Tensor[T]
-
Input:
- 2 tensors of shape Nb observations, features Note: contrary to Numpy default each row is an observations while echo column represent a feature/variable observed.
Returns:
- The unbiased covariance (normalized by the number of observations - 1) in the shape features, features