New methods for modelling real-world extremes. This project aims to develop new theory and methods for analysing and predicting extreme values observed in real-world processes. Many existing techniques are limited by convenient mathematical assumptions that commonly do not hold in practice: dependence at asymptotic levels, process stationarity, and that the observed data are direct measurements of the process of interest. As a result, using these techniques may produce undesirable results. Expec ....New methods for modelling real-world extremes. This project aims to develop new theory and methods for analysing and predicting extreme values observed in real-world processes. Many existing techniques are limited by convenient mathematical assumptions that commonly do not hold in practice: dependence at asymptotic levels, process stationarity, and that the observed data are direct measurements of the process of interest. As a result, using these techniques may produce undesirable results. Expected outcomes of this project include theoretically justified data analysis techniques that can accurately model extreme values seen in the real world. Project benefits include more realistic analyses of nationally important applications in climate, bushfire insurance risk, and anomaly detection.Read moreRead less
Loss-based Bayesian Prediction. This project proposes a new paradigm for prediction. Using state-of-the-art computational methods, the project aims to produce accurate, fit for purpose, predictions which, by design, reduce the loss incurred when the prediction is inaccurate. Theoretical validation of the new predictive method, without reliance on knowledge of the correct statistical model, is an expected outcome, as is an extensive numerical assessment of its performance in empirical settings. T ....Loss-based Bayesian Prediction. This project proposes a new paradigm for prediction. Using state-of-the-art computational methods, the project aims to produce accurate, fit for purpose, predictions which, by design, reduce the loss incurred when the prediction is inaccurate. Theoretical validation of the new predictive method, without reliance on knowledge of the correct statistical model, is an expected outcome, as is an extensive numerical assessment of its performance in empirical settings. The new paradigm should produce significant benefits for all fields in which the consequences of predictive inaccuracy are severe. Problems that lead to substantial economic, financial or environmental loss if predictions are incorrect will be given particular attention.Read moreRead less
The validation of approximate Bayesian computation. This project aims to establish the theoretical validity of approximate Bayesian computation (ABC) and to develop diagnostic methods for assessing its reliability in empirical applications. Given the increased complexity of modern statistical models, new ways of conducting statistical inference are needed. Approximate Bayesian computation is a new statistical tool. This project expects its findings will be useful in all fields where complex phen ....The validation of approximate Bayesian computation. This project aims to establish the theoretical validity of approximate Bayesian computation (ABC) and to develop diagnostic methods for assessing its reliability in empirical applications. Given the increased complexity of modern statistical models, new ways of conducting statistical inference are needed. Approximate Bayesian computation is a new statistical tool. This project expects its findings will be useful in all fields where complex phenomena feature and approximate methods are the only feasible way of understanding those phenomena.Read moreRead less
Semi-parametric bootstrap-based inference in long-memory models. Given the long lead times involved in implementing economic decisions, a clear understanding of the long-term dynamics driving key variables is crucial. This project will produce significant advances in the analysis of long-range dependence, with decisions underpinned by more accurate and robust statistical information as a consequence.
Approximate Bayesian computation in state space models. Economic and financial data frequently exhibit dynamic patterns, driven by unobserved processes that relate to the behaviour of economic agents, or to institutional and technological change. To gain insight into such 'latent' processes is of paramount importance in terms of both understanding the economy and producing accurate, readily up-dated, forecasts of its future performance. Using a Bayesian approach, new simulation-based statistical ....Approximate Bayesian computation in state space models. Economic and financial data frequently exhibit dynamic patterns, driven by unobserved processes that relate to the behaviour of economic agents, or to institutional and technological change. To gain insight into such 'latent' processes is of paramount importance in terms of both understanding the economy and producing accurate, readily up-dated, forecasts of its future performance. Using a Bayesian approach, new simulation-based statistical methods for analysing latent variable models are proposed. Emphasis is given to the development of relatively simple techniques that are applicable to a wide range of empirically relevant models, with a view to improving the access of non-specialists to this powerful form of statistical analysis.Read moreRead less