svdUnfold is a python implementation of singular value decomposition data unfolding based on SVD Approach to Data Unfolding.
import svdunfold
unfold = svdunfold.SVDunfold(x_ini, b, A, B)x_ini is a numpy histogram with the initial generated distribution, representing your expectation for the true underlying distribution. A is a numpy array of the response matrix describing the detector effect, e.g. smearing and efficiency. b is a numpy histogram containing the true measured data distribution and B is a numpy array of the covariance matrix of the measurement.
unfold.transform_system()
d = unfold.get_abs_d()transform_system() prepares the system for unfolding, calculating the decomposition coefficients d. After choosing an appropriate value for the regularization parameter k, simply do:
unfold.unfold(k)The unfolded distribution and the corresponding covariance matrix can be retrieved with:
x_unfolded = unfold.get_unfolded_distribution()
X_covariance = unfold.get_unfolded_cov_matrix()See this notebook for an example.