摘要: |
The Office of Aerospace Medicine needs the ability to define aeromedical hazards and characterize risk, as informed by all relevant data sources, during pilot medical certification in a manner that aligns with the FAA’s approach to safety risk management. Existing aeromedical data produced from the FAA’s pilot medical certification business processes is limited in terms of scope, quantity, and quality, which has stymied efforts to calculate relatively unbiased risk estimators for aeromedical hazards. One approach to mitigate this situation is to use alternative data sources, such as commercial and government-generated healthcare datasets, to calculate risk estimators for aeromedical hazards such This project will seek to identify the best algorithms for forecasting a pilot’s aeromedical risk using commercial healthcare data. The project will develop, apply, and compare several algorithms for forecasting a pilot’s aeromedical risk using commercial healthcare datasets. Additionally, it will provide recommendations for using model performance metrics to choose the best algorithm for forecasting pilot aeromedical risk. |