Faculty
Gregory Cooper, MD, PhD
Research Interests
- Application of decision theory, probability theory, Bayesian statistics, and artificial intelligence to biomedical informatics research problems
- Causal modeling and discovery from clinical and high-throughput molecular data
- Computer-aided medical diagnosis and prediction
- Machine-learning approaches to improving patient safety
- Biosurveillance of disease outbreaks
Appointments and Positions
Current Research Projects and Collaborations
Dr. Cooper’s past and current research involves the application of decision theory, probability theory, machine learning, Bayesian statistics, and artificial intelligence to biomedical informatics research problems. He has been investigating those topic areas for the past 25 years and has published over 110 peer-reviewed papers. He is currently involved in the following research projects:
Predicting Patients’ Outcomes from Clinical and Genome-Wide Data (PI of an R01 grant from the National Library of Medicine, NIH): The primary goals of this project are to develop, implement, and evaluate computer-based models that predict patient outcomes based on clinical and genome-wide data.
Bayesian Detection and Characterization of Disease Outbreaks (PI of a research project within a P01 grant from the CDC): The goals of this project are to develop, implement, deploy, and evaluate a Bayesian system for detecting and characterizing the outbreak of diseases based on input from a disease case detection system, which is a companion P01 project.
Discovering Complex Anomalous Patterns in Data (PI of a Pitt subcontract of an NSF grant to CMU): The goals of this project are to develop, implement, and evaluate a general and widely applicable framework for detecting potentially complex statistical patterns from data about entities in some set of interest, such as patterns of maintenance of jet aircraft in a fleet.
Detecting Deviations in Clinical Care in ICU Data Streams (Co-investigator of an R01 from NIGMS, NIH): The goals of this project are to develop, implement, and evaluate computer-based methods that model usual clinical care and then apply those models to detect individual patient care that is anomalous. Within the ICU domain this project will investigate how often such anomalies correspond to medical management errors.
Recent Publications
Neill DB, Cooper GF. A multivariate Bayesian scan statistics for early event detection and characterization. Accepted by Machine Learning (to appear).
Shen Y, Cooper GF. A New Prior for Bayesian Anomaly Detection – Application to Biosurveillance. Accepted by Methods of Information in Medicine July 2009 (to appear).
Jiang X, Wallstrom G, Cooper GF, Wagner MM. Bayesian prediction of an epidemic curve. Journal of Biomedical Informatics; Feb 42(1):90-9 (2009). PMID: 18593605.
Jiang X, Cooper GF. A Bayesian Network Model for Spatial Event Surveillance. Accepted by International Journal of Approximate Reasoning January 2009 (to appear).
Shen Y, Adamou C, Dowling JN, Cooper GF. Estimating the joint disease outbreak-detection time when an automated biosurveillance system is augmenting traditional clinical case finding. Journal of Biomedical Informatics; 41:224-231 (2008). PMID: 18194876.
Dara J, Dowling JN, Travers D, Cooper GF, Chapman WW. Evaluation of preprocessing techniques for chief complaint classification, Journal of Biomedical Informatics; 41(4):612-23 (2008). PMID: 18166502.
Sutovsky P and Cooper GF. Hierarchical explanation of inference in Bayesian
networks that represent a population of independent agents. In: Proceedings of the 18th European Conference on Artificial Intelligence (2008) 214-218.
Valko M, Cooper G, Seybert A, Visweswaran S, Hauskrecht M. Conditional Anomaly Detection Methods for Patient-Management Alert Systems. In: Proceedings of the Workshop on Machine Learning in Health Care Applications of the 25th International Conference on Machine Learning(2008).