Anymatrix is a MATLAB toolbox that provides an extensible collection of matrices, organized in groups, with the ability to search the collection by matrix properties.

Currently, the built-in collection contains 146 matrices organized into 7 groups:

• contest - the CONTEST test matrix toolbox of random matrices from networks.
• core - miscellaneous matrices.
• gallery - matrices from the MATLAB gallery.
• hadamard - a large collection of Hadamard matrices.
• matlab - other MATLAB matrices (not in gallery).
• nessie - matrices from real-life networks.
• regtools - matrices from regularization problems.

List the groups in the collection.

G = anymatrix('groups')

Show matrices in the core group.

M = anymatrix('groups', 'core')

Show help of a beta matrix in the core group.

anymatrix('help', 'core/beta')

Show properties appended to the beta matrix.

P = anymatrix('properties', 'core/beta')

Generate the 10-by-10 beta matrix.

[A, R] = anymatrix('core/beta', 10)

Start by listing all the supported properties by the anymatrix.

P = anymatrix('properties')

Search for matrices that do not have a property 'symmetric'.

M = anymatrix('properties', 'not symmetric')

Look for matrices that are tridiagonal and not positive, or tridiagonal and symmetric.

M = anymatrix('properties', 'tridiagonal and (symmetric or not positive)')

Look for matrices that mention ' zero ' in their help comments.

M = anymatrix('lookfor', ' zero ')

For invocations that use only character vectors or strings, the MATLAB command form of invocation (as opposed to the functional form) can be used:

anymatrix help core/beta

Commands can be abbreviated to the first one or two letters that uniquely specify them:

anymatrix p core/beta

List all the available commands and shorthand aliases of commands.

anymatrix('help')

The following groups are available for downloading into Anymatrix.

• HPL-AI-Matrix: matrix designed for use in the HPL-AI Benchmark.

• matrices-correlation-invalid: collection of invalid correlation matrices.

• matrices-mp-cosm: collection of matrices for testing the matrix cosine.

• randsvdfast-matlab: provides functionality similar to anymatrix('gallery/randsvd') but uses a faster algorithm.

MATLAB 2020b or newer is required.

Nicholas J. Higham and Mantas Mikaitis. Anymatrix: An Extendable MATLAB Matrix Collection, Users' Guide. MIMS EPrint 2021.15, Manchester Institute for Mathematical Sciences, The University of Manchester, UK. Oct. 2021.

Nicholas J. Higham and Mantas Mikaitis. Anymatrix: An Extendable MATLAB Matrix Collection. MIMS EPrint 2021.16, Manchester Institute for Mathematical Sciences, The University of Manchester, UK. Oct. 2021.

The code is distributed under the terms of the BSD 2-Clause License; see license.txt.