Statistics and Mathematics colloquium: Performance of Model Averaged Tail Area Confidence Interval and Confidence Intervals That Utilize Uncertain Prior Information
You are welcome to attend the following joint Statistics and Mathematics colloquium (part of the Colloquium Series of the Department of Mathematical and Physical Sciences) at La Trobe University organised together with the PVSeminar
- Thursday 27 July 2023 12:00 pm until Thursday 27 July 2023 01:00 pm (Add to calendar)
- Andriy Olenko
- Presented by:
- Ayesha Perera
- Type of Event:
Every model has uncertainties in the variables that it should include.
The Model Averaged Tail Area (MATA) confidence interval has been proposed as a way to deal with model uncertainty. In the Scandinavian Journal of Statistics in 2016, Kabaila, Welsh, and Abeysekara considered the case of two nested linear regression models to assess the performance of MATA confidence intervals.
We extend this study to include the case of three nested linear models.
Additionally, we examine the study conducted by Kabaila & Giri in 2009, where they introduced computationally convenient formulas to compute the confidence interval of two nested linear regression models with uncertain prior information that the simpler model is true. In their computations, they utilized cubic splines.
We aim to improve continuity and computation performance by replacing cubic splines with sinc function interpolations.
Furthermore, we extend this study to the case of three nested linear regression models and derive computationally convenient formulas for the computation of the confidence interval.
Other events by type
- Exploring the cultural phenomenon of the dick pic
22 Sept 2023 6:30 pm
- Maritime Security and Hedging in the South China Sea
26 Sept 2023 11:00 am
- Ideas & Society Program: Disability and Australian Society: How Well Are We Doing?
5 Oct 2023 5:00 pm
- Out In Suburbia: The Untold Story of LGBTIQ+ People in Australia's Outer Suburbs
6 Oct 2023 9:30 am
- John Richards Centre 2023 Oration (Hybrid event)
25 Oct 2023 5:15 pm