propensity-score-matching-thesis

Source code and resources for my MSc Statistics thesis on matching methods for use with observational studies.

Matching is a field within causal inference which attempts to reduce the effects of selection bias on parameter estimates; for example, estimates of the Average Treatment Effect (ATE).

Repository Overview

• paper contains the LaTeX source (and .bib files) for my thesis;
• matching is a python package I have developed for matching observational data;
• experiments contains experiments conducted using matching; run the experiments via

python -m experiments

• docs contains extensive documentation for the package matching

Development Information

• Install rqeuirements via python -m pip install -r requirements.txt
• Python package written in python 3.8.6
• Docstrings follow numpydoc style, and makes exhaustive use of Python's gradual typing library