These files are the simulation codes used in the article "Deconvolution of repeated measurements corrupted by unknown noise", by Jérémie Capitao-Miniconi, Elisabeth Gassiat and Luc Lehéricy.

They were implemented in Python 3.12.7 with scipy 1.13.1.
Date: 15/05/2025.

Description of the files:
- auxiliary functions.py gathers all functions needed by the other scripts. It is not needed to open it to run simulations.
- compute_estimators.py is the first step of the estimation. It computes an estimator per set of parameters (m, nu, h), but does not perform any selection. It generates save files in sub-folders, one per case and number of observations, which will be used by the two following scripts.
- hold_out.py implements the data-driven selection procedure described in Section 6.4 and generates the plots of Figure 5.
- synthesis_oracle.py identifies the oracle set of parameters and its risk and generates the plots of Figure 1-4.