“Estimating customer service rates, loan defaults, and cancer survival: Some applications of a new hazard regression method” by Dr. Donald Lee
Dr. Donald Lee
Associate Professor of Operations
Yale School of Management
Hazard rate models permeate many fields of scientific inquiry. In this talk we will survey some modern applications to queueing, fintech, marketing, and early warning systems for adverse medical events. Central to these applications is the need to estimate the underlying hazard function using as few assumptions as possible. By fusing machine learning with survival analysis, we propose a flexible estimation method that generalizes all existing (semi)parametric models. To illustrate the benefit of the flexibility afforded by our method, we use it to resolve an on-going debate in healthcare operations regarding the relationship between service rate and workload at an emergency department.
Based on joint work with Ningyuan Chen (HKUST)