Showing 1–2 of 2 results for author: Austin, M
-
Heavy tails and negative correlation in a binomial model for sports matches: applications to curling
Authors:
John Fry,
Mark Austin,
Silvio Fanzon
Abstract:
A binomial model for sports matches is developed making use of the maximum possible score $n$ in a game. In contrast to previous approaches the scores of the two teams are negatively correlated, abstracting from a scenario whereby teams cancel each other out. When $n$ is known, analytical results are possible via a Gaussian approximation. Model calibration is obtained via generalized linear modell…
▽ More
A binomial model for sports matches is developed making use of the maximum possible score $n$ in a game. In contrast to previous approaches the scores of the two teams are negatively correlated, abstracting from a scenario whereby teams cancel each other out. When $n$ is known, analytical results are possible via a Gaussian approximation. Model calibration is obtained via generalized linear modelling, enabling elementary econometric and strategic analysis to be performed. Inter alia this includes quantifying the Last Stone First End effect, analogous to the home-field advantage found in conventional sports. When $n$ is unknown the model behaviour is richer and leads to heavy-tailed non-Gaussian behaviour. We present an approximate analysis of this case based on the Variance Gamma distribution.
△ Less
Submitted 18 June, 2024;
originally announced June 2024.
-
Modeling restricted enrollment and optimal cost-efficient design in multicenter clinical trials
Authors:
Vladimir Anisimov,
Matthew Austin
Abstract:
Design and forecasting of patient enrollment is among the greatest challenges that the clinical research enterprize faces today, as inefficient enrollment can be a major cause of drug development delays. Therefore, the development of the innovative statistical and artificial intelligence technologies for improving the efficiency of clinical trials operation are of the imperative need. This paper i…
▽ More
Design and forecasting of patient enrollment is among the greatest challenges that the clinical research enterprize faces today, as inefficient enrollment can be a major cause of drug development delays. Therefore, the development of the innovative statistical and artificial intelligence technologies for improving the efficiency of clinical trials operation are of the imperative need. This paper is describing further developments in the innovative statistical methodology for modeling and forecasting patient enrollment. The underlying technique uses a Poisson-gamma enrollment model developed by Anisimov & Fedorov in the previous publications and is extended here to analytic modeling of the enrollment on country/region level. A new analytic technique based on the approximation of the enrollment process in country/region by a Poisson-gamma process with aggregated parameters is developed. Another innovative direction is the development of the analytic technique for modeling the enrollment under some restrictions (enrollment caps in countries). Some discussion on using historic trials for better prediction of the enrollment in the new trials is provided. These results are used for solving the problem of optimal trial cost-efficient enrollment design: find an optimal allocation of sites/countries that minimizes the global trial cost given that the probability to reach an enrollment target in time is no less than some prescribed probability. Different techniques to find an optimal solution for high dimensional optimization problem for the cases of unrestricted and restricted enrollment and for a small and large number of countries are discussed.
△ Less
Submitted 25 December, 2022;
originally announced December 2022.
