Heidi Gurung
Heidi Gurung
Data Modeler at IEM
Dr. Gurung studied Bioinformatics with research in the application of machine learning to DNA analysis, epidemiological models, and infectious diseases. Experience in development and research of machine learning approaches to regression and classification problems in various fields, including epidemiology and sound (music). Highly effective technical team leader with 4 years of experience delivering numerous successful projects through collaboration with external teams and government personnel. Dr. Gurung is an experienced software engineer with over 18 years of software development, system administration, and software system troubleshooting experience, including software design, implementation, and evaluation.
Track: Accounting and Finance
A Toolbox Approach to COVID Fraud Detection
In this presentation we will discuss the different tools used when analyzing fraud in unemployment claims due to COVID. We will consider basic discovery analytics, social network analysis, and machine learning techniques for identifying and reporting potential fraud. We will also discuss lessons learned regarding working with large, messy datasets (more than 25 million rows and around 100 columns), including cleaning, de-duplicating, restructuring, etc. as needed for effective analysis. Finally, we will look briefly at automating query and reporting activities for consistent, reproducible results and versioning and storage of scripts for knowledge and capability transfer between development and maintenance teams. This talk will cover the start-to-finish of an investigation project for unemployment fraud detection.
