6 papers
Design-Based Supervised Learning with Noisy Human Labels
Robert Chew, Matthew R. Williams
Researchers increasingly use automated classifiers to label unstructured data for statistical analysis. Existing rectification methods can correct errors in these automated labels…
Design Effect Ratios for Bayesian Survey Models: A Diagnostic Framework for Identifying Survey-Sensitive Parameters
JoonHo Lee, Matthew R. Williams, Terrance D. Savitsky
Bayesian hierarchical models are increasingly fitted to complex survey data through weighted pseudo-posteriors, with a post-processing step that rescales the posterior to match a d…
Privacy Amplification for Synthetic data using Range Restriction
Jingchen Hu, Matthew R. Williams, Terrance D. Savitsky
We introduce a new class of range restricted formal data privacy standards that condition on owner beliefs about sensitive data ranges. By incorporating this additional information…
Bayesian Pseudo Posterior Mechanism for Differentially Private Machine Learning
Robert Chew, Matthew R. Williams, Elan A. Segarra +3
Differential privacy (DP) is becoming increasingly important for deployed machine learning applications because it provides strong guarantees for protecting the privacy of individu…
Representative dietary behavior patterns and associations with cardiometabolic outcomes in Puerto Rico using a Bayesian latent class analysis for non-probability samples
Stephanie M. Wu, Abrania Marrero, Matthew R. Williams +4
There is limited understanding of how dietary behaviors cluster together and influence cardiometabolic health at a population level in Puerto Rico. Data availability is scarce, par…
Thresholding Nonprobability Units in Combined Data for Efficient Domain Estimation
Terrance D. Savitsky, Matthew R. Williams, Julie Gerrshunskaya +1
Quasi-randomization approaches estimate latent participation probabilities for units from a nonprobability / convenience sample. Estimation of participation probabilities for conve…