Polynomial approximations for computer models and data analytics.

Uncertainty quantification

Polynomial chaos with independent and correlated input parameters using tensor and sparse grids, least squares, or compressive sensing techniques.

Dimension reduction

Data-driven dimension reduction using active subspaces and polynomial ridge approximations; Sobol' indices for understanding parameter importance.

Machine learning

Piecewise polynomial regression using trees along with functions for estimating posterior variance. Optimisation using trust-region based polynomial surrogates.


Effective Quadratures is a collection of utilities for understanding the input-output nature of models and data. Our tools are powered by our open-source code equadratures. The latest version of the code is Narwhal v9.0. To download and install the code, please use the python package index command:

pip install equadratures

To set the code up, try the following:

from equadratures import *
import numpy as np

def my_model(x):
	return x**2 + 0.2 * x - np.sin(x)

parameter = Parameter(distribution='uniform', lower=-1., upper=1., order=3)
basis = Basis('univariate')
poly = Poly(parameter, basis, 'numerical-integration')