Fletcher Lu

Fletcher Lu,  Assistant Professor
PhD (Computer Science), University of Waterloo

Personal Profile
During his graduate studies at the  University of Waterloo, he was a member of the Waterloo Institutes of Health Informatics Research group there. He obtained a Computer Science PhD degree from the University of Waterloo in the fall of 2003.  Much of my research directly relates to data mining with health applications.  I specifically have published work in the detection of fraud in healthcare insurance.  I have forged industrial research collaborations with health insurance businesses such as Manulife Financial.  I have obtained external funding resources from the Canadian Institute of Health Research for research into healthcare fraud.  I also obtained external funding from the Natural Science and Engineering Research Council of Canada in the form of a post-doctoral research fellowship of $52,000/year (annually renewable up to two years) awarded in October 2004.

Research Areas:
By applying machine learning data mining techniques to detecting fraud in both business and health areas and compare them to traditional artificial intelligence approaches such as neural networks.This project is principally application oriented, exploring using reinforcement learning techniques and comparing them with previous approaches to fraud detection.  During his graduate work he looked at Bayesian Probability approaches to performing medical diagnosis. He analyzed various diagnostic systems that typically dealt with various types of cancers and attempted to compare the computer diagnosis and treatment with human diagnosis and treatment methods. His current work involves development of a fraud auditing algorithm applied to healthcare insurance data.

His work best falls under the categories of Informatics of Health Institutions and Data Mining as much of it deals with searching vast amounts of data typically from large institutions such as healthcare units or insurance companies.

Contact Information
Office:     N/A
Phone:   
N/A
Email:     N/A