LowRank.ProportionalDecomposer

LowRank.ProportionalDecomposer#

LowRank.ProportionalDecomposer.py

This module contains the functions to decompose a curve according to specified peak area proportions.

compute_egh_area_fast(h, sigma, tau)#

Compute the area under an exponentially modified Gaussian (EMG) function quickly using an analytical formula.

Parameters:

  • h (float) – The height of the EMG function.

  • sigma (float) – The standard deviation of the Gaussian component.

  • tau (float) – The decay constant of the exponential component.

Returns:

The area under the EMG function.

Return type:

float

decompose_icurve_proportionally(x, y, decompargs, **kwargs)#

Decompose the given curve (x, y) into a sum of Exponential- Gaussian Hybrid (EGH) functions, with area proportions specified in decompargs[‘proportions’]. The number of peaks is determined by the length of the proportions list. The initial peak positions are estimated based on the cumulative area of the smoothed curve

Parameters:

  • debug (bool, optional) – If True, enables debug mode with additional output.

  • proportions (list of float) – The area proportions for each peak.

  • tau_limit (float, optional) – The maximum allowed ratio of tau/sigma for each peak. Default is 0.6.

  • max_sigma (float, optional) – The maximum allowed sigma for each peak. Default is 17.

  • dev_weights (tuple of float, optional) – Weights for the deviation penalties in the optimization.

  • basinhopping (bool, optional) – If True, use basinhopping optimization. Default is False. This can help to escape local minima.

Returns:

params – A 2D array of shape (number of peaks, 4), where each row contains the parameters (height, mean, sigma, tau) of the corresponding EMG function.

Return type:

np.ndarray

Contents