pastis.optimization.mds.estimate_X

pastis.optimization.mds.estimate_X(counts, alpha=- 3.0, beta=1.0, ini=None, verbose=0, use_zero_entries=False, precompute_distances=False, bias=None, random_state=None, type='MDS2', factr=1000000000000.0, maxiter=10000)

Estimating the structure from contact count data using MDS/NMDS

Parameters
countsndarray (n, n)

The (normalized) contact count matrix as a dense ndarray

alphafloat, optional, default: -3

Coefficient of the power law used in converting interaction counts to wish distances

betafloat, optional, default: 1.

Coefficient corresponding to the scaling factor of the structure.

ini{None, “random”, ndarray (n, 3)}, optional, default: None

The initial point of the optimization. If None or “random”, will be randomly set. Else, should be an ndarray of shape (n, 3) where n is the number of bins in the contact count matrix.

use_zero_entriesboolean, optional, default: False

Whether to use contact counts that are 0.

precompute_distanceboolean, optional, default: False

If provided, then counts is considered to be the wish distances to use in the MDS optimization.

random_stateint, RandomState instance, default=None

Determines random number generation. Use an int to make the randomness deterministic.

type{“MDS1”, “MDS2”}, optional, default: “MDS2”

Whether to apply MDS1 or “MDS2” (weighted MDS). See companyon article for more information.

factrfloat, optional, default: 1e12

The precision of the optimization algorthim. See scipy.optimize.fmin_l_bfgs_b for more information. The higher, the more precise. By default, is set to very high precision.

maxiterinteger, optional, default: 10000

The maximum number of iteration.

Returns
The structure as an ndarray of shape (n, 3).

Examples using pastis.optimization.mds.estimate_X