Examining the performance of weighting methods under different populations
Principal Investigator
Name
Lei Song
Degrees
PhD
Institution
George Washington University
Position Title
student
Email
About this CDAS Project
Study
PLCO
(Learn more about this study)
Project ID
PLCO-1480
Initial CDAS Request Approval
Feb 15, 2024
Title
Examining the performance of weighting methods under different populations
Summary
When transporting a model to a new cohort, researcher usually would apply weighting methods to the previous models since the distribution might be different. A novel reweighting method is designed to deal with extreme weights and insufficient overlap and thus can help achieve more accurate estimation. We want to examine its performance under the real world scenarios where NLST and PLCO have very distinct population distribution. First the difference is measured to create subgroups of datasets with a gradual change in degree of overlap. Then we would test the performance of various models with weighting methods under complex data structure.
Aims
1. Examine the population distribution difference between PLCO and NLST with numeric measurement.
2. Compare the performance different prediction and weighting methods based on different subgroup of datasets.
3. Test the robustness of the weighting methods under different conditions.
Collaborators
Dr. Qing Pan, from George Washington University
Dr. Chen Hu, from John Hopkins University