nws_tools.generate_randnws¶

nws_tools.generate_randnws(nw, M, method='auto', rwr=5, rwr_max=10)[source]¶

Generate random networks given a(n) (un)signed (un)weighted (un)directed input network

Parameters:

Parameters:	nw : NumPy 2darray Connection matrix of input network M : int Number of random networks to generate (> 1) method : str String specifying which method to use to randomize the input network. Currently supported options are ‘auto’ (default), ‘null_model_und_sign’, ‘randmio_und’, ‘randmio_und_connected’, ‘null_model_dir_sign’, ‘randmio_dir’, ‘randmio_dir_connected’, ‘randmio_und_signed’, ‘randmio_dir_signed’, If method = ‘auto’ then a randomization strategy is chosen based the the properties of the input network (directedness, edge-density, sign of edge weights). In case of very dense networks (density > 75%) the null_model routines are used to at least shuffle the input network’s edge weights. rwr : int Number of approximate rewirings per edge (default: 5). rwr_max : int Maximal number of rewirings per edge to enforce randomization (default: 10). Note that rwr_max has to be greater or equals rwr.
Returns:	rnws : NumPy 3darray Random networks based on input graph nw. Format is rnws.shape = (N,N,M) such that rnws[:,:,m] = m-th N x N random network

nw : NumPy 2darray

Connection matrix of input network

M : int

Number of random networks to generate (> 1)

method : str

String specifying which method to use to randomize the input network. Currently supported options are ‘auto’ (default), ‘null_model_und_sign’, ‘randmio_und’, ‘randmio_und_connected’, ‘null_model_dir_sign’, ‘randmio_dir’, ‘randmio_dir_connected’, ‘randmio_und_signed’, ‘randmio_dir_signed’, If method = ‘auto’ then a randomization strategy is chosen based the the properties of the input network (directedness, edge-density, sign of edge weights). In case of very dense networks (density > 75%) the null_model routines are used to at least shuffle the input network’s edge weights.

rwr : int

Number of approximate rewirings per edge (default: 5).

rwr_max : int

Maximal number of rewirings per edge to enforce randomization (default: 10). Note that rwr_max has to be greater or equals rwr.

Returns:

rnws : NumPy 3darray

Random networks based on input graph nw. Format is rnws.shape = (N,N,M) such that rnws[:,:,m] = m-th N x N random network