Date of Award

Spring 2022

Document Type


Degree Name

Doctor of Philosophy (PhD)


Computer Science

First Advisor

Ahamed, Sheikh Iqbal

Second Advisor

Adibuzzaman, Mohammad

Third Advisor

Madiraju, Praveen


In the era of big data, researchers have access to large healthcare datasets collected over a long period. These datasets hold valuable information, frequently investigated using traditional Machine Learning algorithms or Neural Networks. These algorithms perform great in finding patterns out of datasets (as a predictive machine); however, the models lack extensive interpretability to be used in the healthcare sector (as an explainable machine). Without exploring underlying causal relationships, the algorithms fail to explain their reasoning. Causal Inference, a relatively newer branch of Artificial Intelligence, deals with interpretability and portrays causal relationships in data through graphical models. It explores the issue of causality and works towards an explainability of underlying causal models deeply buried in data. For this dissertation work, the research goal is to use Causal Inference to build an applied framework that lets researchers leverage observational datasets in understanding causal relationships between features. To achieve that, we focus on specific objectives such as (a) the addition of background knowledge to causal structure learning algorithms, (b) the proposal of new causal inference methodologies, (c) generation of theories connecting causality to standard statistical analyses (e.g., Odds Ratio, Survival Analysis), and (d) application of proposed approaches in real-world healthcare problems. This dissertation encapsulates the tasks mentioned above, through various new methodologies and experiments under the rubric of Structural Theory of Causation. We discuss the common research theme in causal inference, historical development, the structural theory of causation, and underlying assumptions. Finally, we explore the impact of these proposed methodologies in real-world treatment controversy of Delirium patients, by examining the efficacy of antipsychotic drugs prescribed in treating Delirium in the ICU, from a curated observational healthcare dataset.