Internal Numpy Type¶

Many of the python bindings accept both scalars and values that can be converted to numpy arrays. The latter are first converted to a specific numpy type, depending on how this package was compiled.

dcprogs.internal_dtype¶

Numpy type used internally

If the package is configured and compiled with -DCPROGS_LONG_DOUBLE (see wiki), then long doubles are used through out. The exact size of long doubles is platform dependent and is typically 80-bit on Linux and OSX and 64-bit on Windows. See Wikipedia for more information. Otherwise, the package defaults to garden variety 64-bit doubles.

Linear Algebra¶

Numpy only provides a subset of its utilities for arrays consisting of long doubles. As a result, this package exposes some of Eigen‘s capabilities, as needed. Their interface is generally reminiscent of the numpy utility they mirror.

dcprogs.likelihood.eig(*args)[source]¶

Computes eigenvalues of a matrix

dcprogs.likelihood.inv(*args)[source]¶

Computes inverse of a non-singular matrix

dcprogs.likelihood.svd()¶

dcprogs.likelihood.det(*args)[source]¶

Computes determinant of a matrix

dcprogs.likelihood.expm(*args)[source]¶

Computes singular value decomposition of a matrix