JOIN US on March 16, 2022 for a special CPNM Research Seminar Series event, featuring Dr. Anthony Kassekert, Chief Analytics Officer, Hazard Identification and Reduction at the U.S. Consumer Product Safety Commission.
Presentation Abstract: Covid 19 has amplified the existing concern that missing information in government data is a major hinderance in determining public policy impacts regarding equity and the impacts of public policy on historically underserved communities. The discrepancies caused by missing values increase the likelihood that hazard rates for underserved populations will be misleading if based solely on the raw data. Using data from the U.S. Consumer Product Safety Commission (CPSC), this paper employs a novel combination of imputation techniques to evaluate how threats to at risk groups changes after missing data is corrected. In particular, the National Electronic Injury Surveillance System (NEISS) data will be used to address risk to vulnerable groups in response to Executive Order 13985 “Advancing Racial Equity and Support for Underserved Communities Through the Federal Government”
This research applies several innovative imputation techniques to mitigate the impact of missing data on emerging hazard detection. The health narrative with each case is analyzed using machine learning to pull information that may improve the prediction of demographic data. The data set is then sent through Bayesian algorithms to predict key indicators of underserved population membership. The results of the imputed model versus the actuals with missing data removed are compared to demonstrate the improvement in assessing risks to vulnerable populations through statistical imputation.
Join US Downtown or on Zoom!
Read More