QuantileInterpolator

class pyirf.interpolation.QuantileInterpolator(grid_points, bin_edges, binned_pdf, quantile_resolution=0.001, normalization=PDFNormalization.AREA)[source]

Bases: DiscretePDFInterpolator

Interpolator class providing quantile interpoalation.

Methods Summary

__call__(target_point)

Providing a common __call__ interface

interpolate(target_point)

Takes a grid of binned pdfs for a bunch of different parameters and interpolates it to given value of those parameters.

Methods Documentation

__call__(target_point)

Providing a common __call__ interface

Parameters:
target_point: np.ndarray, shape=(1, n_dims)

Target for inter-/extrapolation When target_point is outside of the grids convex hull but extrapolator is None

Returns:
Interpolated result.
interpolate(target_point)[source]

Takes a grid of binned pdfs for a bunch of different parameters and interpolates it to given value of those parameters. This function provides an adapted version of the quantile interpolation introduced in [1]. Instead of following this method closely, it implements different approaches to the steps shown in Fig. 5 of [1].

Parameters:
target_point: numpy.ndarray, shape=(1, n_dims)

Value for which the interpolation is performed (target point)

Returns:
f_new: numpy.ndarray, shape=(1, …, n_bins)

Interpolated and binned pdf

References

[1]

B. E. Hollister and A. T. Pang (2013). Interpolation of Non-Gaussian Probability Distributions for Ensemble Visualization https://engineering.ucsc.edu/sites/default/files/technical-reports/UCSC-SOE-13-13.pdf