Monday, April 15 2019
12:30pm - 1:30pm
Statistics Seminar
STaRs Seminar by Tian Yu Yen

Do you believe that your parameter really is a random variable? Intro to Data Consistent Inversion and Interactive Discussion about Implications and Applications

Data Consistent Inversion is a new statistical technique that the Uncertainty Quantification group here at CU Denver has been working on for the past few years. Put simply, this method uses prior beliefs and observations of data to produce an "updated" probability distribution for parameters of interest. Though similar in flavor to standard Bayesian statistical methods, the Data Consistent approach updates the prior beliefs with a consistency criteria as opposed to the data-likelihood used in Bayesian analysis. In this interactive presentation, we will first introduce the basic concepts of the Data Consistent method and then discuss how the question, "Is your parameter actually a random variable?" is central to the differences between Bayesian and Data Consistent approaches.

On a personal note: my goal is to keep this introduction short to leave a lot of time for discussion. I hope to get some feedback about what kinds of questions members of the statistical community have about the Data Consistent method and also brainstorm potential applications and topics for future investigation. I hope you will join me!

Download as iCalendar