Semester of Graduation
Summer 2026
Degree Type
Dissertation
Degree Name
Doctor of Philosophy in Data Science and Analytics
Department
School of Data Science and Analytics - College of Computing and Software Engineering
Committee Chair/First Advisor
Herman Ray
Second Advisor
Weiwei Chen
Third Advisor
Jiajing Huang
Fourth Advisor
Gita Taasoobshirazi
Abstract
Constraint-based causal discovery often relies on conditional independence tests to determine basic associations before determining causality. However, conditional independence tests tend to make strong assumptions about the data type, discrete or continuous. For an abundance of studies, this is not practical as many data sources often include discrete and continuous data, associated with various distributions. Although many researchers opt to use discretization methods, we found this approach results in loss of information and reduces the likelihood of relationship discovery. In this paper, we propose a pseudo correlation-like measure to extend Pearson’s partial correlation test to discrete and non-Gaussian data. Our method leverages Efron’s pseudo R2, to obtain a proxy for correlation between data that does not traditionally meet the assumptions of Pearson’s correlation coefficient. This method offers flexibility as Efron’s pseudo R2 can be estimated using supervised learning methods.