Computer code to reproduce the numerical results presented in "Robust probabilistic inference via a constrained transport metric" (Chakraborty A., Bhattacharya A., Pati D., 2023+). Please reach out in [email protected] for any queries. Here, we present a short decription about the scripts in the repository.
(1) Clustering: Find the R scripts to carry out model based clustering in presence of small perturbation of the data genration mechanism (a) based on Standard/Fractional Bayesian proccedure (FractionalPost_Clustering.R) and D-BETEL based methodology (DBETEL_Clustering.R). This can be utilised to re-produce the results presented in Section 3.1 of the paper.
(2) Generalised linear regression: Find the R scripts to carry out robust Poisson regression in presence of small perturbation of the data genration mechanism (a) based on Standard Bayesian proccedure, (b) D-BETEL based methodology (DBETEL_PoissonRegression.R), and (c) Bayesian exponentially tilted empirical likelihood based on moment conditional models (MCM_PoissonRegression.R). This can be utilised to re-produce the results presented in Section 3.2 of the paper.
(3) Demographic Parity: Find the R scripts to carry out D-BETEL based analysis of COMPAS data (DBETEL_DP_COMPAS.R), and DAIC (DBETEL_DP_DAIC.R) data that ensure statistical/demographic parity. This can be utilised to re-produce the results presented in Section 4 of the paper.
DATA AVAILABILITY: The popular COMPAS dataset includes detailed information on criminal history for the defendants in Broward County, Florida, freely available from https://www.propublica.org/datastore/dataset/compas-recidivism-risk-score-data-and-analysis.
The Distress Analysis Interview Corpus (DAIC) is a multi-modal collection of semi-structured clinical interviews, available upon request from https://dcapswoz.ict.usc.edu/.