Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore a research seminar presentation that delves into the challenges and solutions of evaluating probabilistic data linkage through the creation of longitudinal synthetic datasets for population reconstruction. Learn how researchers address the difficulty of evaluating linkage success when true linkage information is unknown, moving beyond the limitations of small hand-linked datasets. Discover the methodologies used to define model behavior, avoid distribution clashes, verify population statistical accuracy, and initialize models to meet temporal requirements. The presentation covers both previous and current modeling approaches, concluding with discussions on linkage evaluation applications and other potential uses for these synthetic datasets in population studies.