deeptime.numeric

The numeric package contains numerical utilities.

General numerical tools

drop_nan_rows(arr, *args)

Remove rows in all inputs for which arr has _np.nan entries.

is_diagonal_matrix(matrix)

Checks whether a provided matrix is a diagonal matrix, i.e., \(A = \mathrm{diag}(a_1,\ldots, a_n)\).

is_square_matrix(arr)

Determines whether an array is a square matrix.

schatten_norm(mat[, order, hermitian])

Computes the r-Schatten norm

Numerical tools for eigenvalue problems

eigs(matrix[, n_eigs, which])

Computes the eigenvalues and eigenvectors of matrix.

eig_corr(C0, Ct[, epsilon, method, ...])

Solve generalized eigenvalue problem with correlation matrices C0 and Ct

sort_eigs(evals, evecs[, order])

Sorts the eigenvalues and eigenvectors by descending norm of the eigenvalues or lexicographically.

spd_eig(W[, epsilon, method, ...])

Rank-reduced eigenvalue decomposition of symmetric positive definite matrix.

spd_truncated_svd(mat[, dim, eps])

Rank-reduced singular value decomposition of symmetric positive (semi-)definite matrix.

spd_inv(W[, epsilon, method])

Compute matrix inverse of symmetric positive-definite matrix \(W\).

spd_inv_split(W[, epsilon, method, ...])

Compute \(W^{-1} = L L^T\) of the symmetric positive-definite matrix \(W\).

spd_inv_sqrt(W[, epsilon, method, return_rank])

Computes \(W^{-1/2}\) of symmetric positive-definite matrix \(W\).

ZeroRankError

Input matrix has rank zero.