|
Activity Number:
|
178
- Contributed Poster Presentations: Royal Statistical Society
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 31, 2017 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Royal Statistical Society
|
|
Abstract #325426
|
|
|
Title:
|
Implementing Monte Carlo Tests with Multiple Thresholds
|
|
Author(s):
|
Georg Hahn* and Axel Gandy and Dong Ding
|
|
Companies:
|
Imperial College London and Imperial College London and Imperial College London
|
|
Keywords:
|
Algorithm ;
Finite time ;
Hypothesis ;
P-value ;
Significance
|
|
Abstract:
|
Software packages usually report the significance of statistical tests using p-values. We are interested in computing the significance of a hypothesis H with respect to several thresholds simultaneously -- with the caveat that the p-value p corresponding to H is unknown and can only be approximated using Monte Carlo simulation. Instead of considering a set of thresholds, this poster presents a more general construction which allows to compute a decision of p with respect to user-specified intervals (called "p-value buckets"): Whereas non-overlapping buckets lead to classical decisions in expected infinite runtime, suitably chosen overlapping buckets allow guaranteed decisions in finite time which are reported in a new fashion that extends the widespread */**/*** significance notation.
|
Authors who are presenting talks have a * after their name.
