sctk._pipeline.fit_gaussian

sctk._pipeline.fit_gaussian(x, n_components=array([1, 2, 3, 4, 5, 6, 7, 8, 9, 10]), threshold=0.05, xmin=None, xmax=None, plot=False, nbins=500, hist_bins=100) → tuple

Fit a Gaussian mixture model to a 1D numpy array.

This function fits a Gaussian mixture model to a 1D numpy array using the specified number of components and threshold. It returns the lower and upper bounds of the fitted Gaussian distribution, as well as the fitted Gaussian mixture model.

Parameters:

  • x – 1D numpy array to fit a Gaussian mixture model to.

  • n_components – Number of components to use for the Gaussian mixture model. The best GMM will be selected based on BIC.

  • threshold – Threshold value for determining the lower and upper bounds of the fitted Gaussian distribution.

  • xmin – Minimum value to use for the fitted Gaussian distribution. If not provided, this will be set to the minimum value of x.

  • xmax – Maximum value to use for the fitted Gaussian distribution. If not provided, this will be set to the maximum value of x.

  • plot – If True, plot the fitted Gaussian distribution.

  • nbins – Number of bins to use for the distribution of x.

  • hist_bins – Number of bins to use for the histogram of x in the plot.

Returns:

Tuple containing the lower bound, upper bound, and Gaussian mixture model.

Raises:

None. –

Examples

>>> import numpy as np
>>> from sctk import fit_gaussian
>>> x = np.random.normal(loc=0, scale=1, size=1000)
>>> x_left, x_right, gmm = fit_gaussian(x, n_components=[2], threshold=0.1)