Researchers at Brown University have developed software designed to eliminate multiple hypothesis testing errors in interactive data exploration and visualization systems using real-time statistical safeguards.
Their QUDE system, unveiled last week at the ACM SIGMOD/PODS Conference in Chicago, is designed so researchers can monitor the risk of false discovery as hypothesis tests are ongoing.
"The idea is that you have a budget of how much false discovery risk you can take, and we update that budget in real time as a user interacts with the data," says Brown professor Eli Upfal. "We also take into account the ways in which a user might explore the data. By understanding the sequence of their questions, we can adapt our algorithm and change the way we allocate the budget."
Brown professor Tim Kraska says the QUDE software presents statistical significance in the form of color-coded feedback, with green signaling a significant finding and red representing higher statistical uncertainty.
From News from Brown
View Full Article
Abstracts Copyright © 2017 Information Inc., Bethesda, Maryland, USA