A0917
Title: Robust group testing for prevalence estimation against uncertain test error mechanisms
Authors: Shih-Hao Huang - National Central University (Taiwan) [presenting]
Abstract: In group testing applications, a primary focus is to estimate the prevalence of a trait in the presence of testing errors. However, the incorporation of testing error models may introduce substantial bias into prevalence estimation when misspecified and inflates variance when involving additional parameters. To mitigate these challenges, a robust estimation method is proposed within group testing frameworks to alleviate model misspecification bias. Additionally, efficient design algorithms are developed for data collection to complement the estimation technique, thereby enhancing prevalence estimation by reducing variance. The simulation experiments demonstrate that the approaches usually result in reasonably small mean squared errors compared to conventional ones.
