Innovations in Bayesian likelihood-free inference. Bayesian inference is a statistical method of choice in applied science. This project will develop innovative tools which permit Bayesian inference in problems considered intractable only a few years ago. These methods will expedite advances in multidisciplinary research across a range of applications. With these foundations, this project will accelerate national research efforts into improving frameworks for projecting trends in water availabil ....Innovations in Bayesian likelihood-free inference. Bayesian inference is a statistical method of choice in applied science. This project will develop innovative tools which permit Bayesian inference in problems considered intractable only a few years ago. These methods will expedite advances in multidisciplinary research across a range of applications. With these foundations, this project will accelerate national research efforts into improving frameworks for projecting trends in water availability and management, the impact of climate extremes, telecommunications engineering, HIV and infectious disease modelling and biostatistics. With many sectors unable to recruit appropriately trained statisticians within Australia, this project will train four PhD students in Bayesian statistics.
Read moreRead less
New methods for small group analysis from sample surveys. National and state averages of statistics on issues such as unemployment, salinity, drought impact, and health often hide large differences between population sub-groups and between small areas. This local variation needs to be understood so that effective policies can be developed and carried out efficiently and their impact monitored. This project will provide, for the first time, robust and efficient methods for providing information o ....New methods for small group analysis from sample surveys. National and state averages of statistics on issues such as unemployment, salinity, drought impact, and health often hide large differences between population sub-groups and between small areas. This local variation needs to be understood so that effective policies can be developed and carried out efficiently and their impact monitored. This project will provide, for the first time, robust and efficient methods for providing information on these variations using data from large-scale national and state surveys. This will lead to significant improvements in the data available for small population groups and small areas, allowing better targeting of policies aimed at addressing local differences.Read moreRead less
Handling Missing Data in Complex Household Surveys. The Australian Bureau of Statistics (ABS) has an extensive program of household surveys that is a key source of information on the social and economic conditions of the population. They provide statistics and data on a large range of social and economic topics, such as health, education, the labour force, income and expenditure. Analysis of household survey data by a variety of organisations underpins policy development and evaluation and the e ....Handling Missing Data in Complex Household Surveys. The Australian Bureau of Statistics (ABS) has an extensive program of household surveys that is a key source of information on the social and economic conditions of the population. They provide statistics and data on a large range of social and economic topics, such as health, education, the labour force, income and expenditure. Analysis of household survey data by a variety of organisations underpins policy development and evaluation and the expenditure of billions of dollars. This project will substantially improve the cost-efficiency and reliability of Australian household survey data, by creating new approaches for handling missing data that deal with the realities of typical household surveys.Read moreRead less
Discovery Early Career Researcher Award - Grant ID: DE130101670
Funder
Australian Research Council
Funding Amount
$370,410.00
Summary
Scalable Bayesian model selection for massive data sets. This project will develop highly innovative, efficient and ultimately effective methodology for Bayesian model selection for large-scale problems which commonly arise in biostatistics and bioinformatics. The resulting methodology will dramatically reduce the duration of analyses in these areas from days or weeks to minutes or hours.
Asymptotic Expansions and Large Deviations in Probability and Statistics: Theory and Applications. Statistics is the major enabling science in a number of disciplines. This is fundamental research in probability and statistics but it has wide applications in Biology and Social Sciences which will ultimately be of national benefit. The behaviour of self normalized sums is an exciting new area of fundamental research that has implications for the application of statistics in many areas. U-statist ....Asymptotic Expansions and Large Deviations in Probability and Statistics: Theory and Applications. Statistics is the major enabling science in a number of disciplines. This is fundamental research in probability and statistics but it has wide applications in Biology and Social Sciences which will ultimately be of national benefit. The behaviour of self normalized sums is an exciting new area of fundamental research that has implications for the application of statistics in many areas. U-statistics for dependent situations has direct application to understanding financial time series and the analysis of sample survey data. Saddlepoint methods provide extremely accurate approximations in a number of important applications.
Read moreRead less
Empirical saddlepoint approximations and self-normalized limit theorems. Finite population sampling and resampling methods such as the bootstrap and randomization methods are central in a number of areas of application and M-estimates are the major method used to give robust methods under mild conditions; in both these areas statistics are used which are Studentized or self-normalized. We will develop asymptotic approaches for such statistics. Saddlepoint and empirical saddlepoint methods will ....Empirical saddlepoint approximations and self-normalized limit theorems. Finite population sampling and resampling methods such as the bootstrap and randomization methods are central in a number of areas of application and M-estimates are the major method used to give robust methods under mild conditions; in both these areas statistics are used which are Studentized or self-normalized. We will develop asymptotic approaches for such statistics. Saddlepoint and empirical saddlepoint methods will be used to give methods which have second order relative accuracy in large deviation regions and we will obtain limit results and Edgeworth approximations. Emphasis will be on obtaining results under weak conditions necessary for applications.Read moreRead less
Semiparametric Regression for Streaming Data. Semiparametric regression converts large and complex data-sets into interpretable summaries from which sound decisions can be made. This project tackles semiparametric regression analysis of streaming data - where the data are so voluminous that they may not be storable in standard computer memory and therefore need to be processed rapidly on arrival and then discarded. Effective solutions necessitate a rethinking of semi-parametric regression and ne ....Semiparametric Regression for Streaming Data. Semiparametric regression converts large and complex data-sets into interpretable summaries from which sound decisions can be made. This project tackles semiparametric regression analysis of streaming data - where the data are so voluminous that they may not be storable in standard computer memory and therefore need to be processed rapidly on arrival and then discarded. Effective solutions necessitate a rethinking of semi-parametric regression and new approaches will be developed. The project will also develop novel theory and methodology for robotics applications. It will allow analysis of streaming and massive data sets that would not be possible using currently available methods, opening up new applications.Read moreRead less
Fast approximate inference methods: new algorithms, applications and theory. This project aims to develop new algorithms and theory for fast approximate inference and lay down infrastructure to aid future extensions. Fast approximate inference methods are a principled and extensible means of fitting large and complex statistical models to big data sets. They come into their own in applications where speed is paramount and traditional approaches are not feasible. The project aims to lead to prac ....Fast approximate inference methods: new algorithms, applications and theory. This project aims to develop new algorithms and theory for fast approximate inference and lay down infrastructure to aid future extensions. Fast approximate inference methods are a principled and extensible means of fitting large and complex statistical models to big data sets. They come into their own in applications where speed is paramount and traditional approaches are not feasible. The project aims to lead to practical outcomes from better business decision-making for insurance data warehouses, to improved medical imaging technology.Read moreRead less
New Bayesian methodology for understanding complex systems using hidden Markov models and expert opinion, environmental, robotics and genomics applications. This project aims to merge four areas of intense international interest in describing complex systems: hidden Markov models and mixtures, semi-parametric and nonparametric approaches, true combination of expert opinion with data, and new Bayesian computational methods based on perfect sampling and particle sampling. The project will signific ....New Bayesian methodology for understanding complex systems using hidden Markov models and expert opinion, environmental, robotics and genomics applications. This project aims to merge four areas of intense international interest in describing complex systems: hidden Markov models and mixtures, semi-parametric and nonparametric approaches, true combination of expert opinion with data, and new Bayesian computational methods based on perfect sampling and particle sampling. The project will significantly contribute to statistical methodology and its ability to inform about real-world problems. A strong focus on applications to genomics, robotics and environmental modelling will bring immediate research and monetary benefit for industry. Expected outcomes include enhanced cross-disciplinary and international linkages, publications, industry-funded projects and highly trained graduates.Read moreRead less
Frontiers in Data Science: Analysing Distributions as Data. This project aims to develop the statistical foundations of a new approach to analysing large and complex data, based on building distributional approximations of the data, which can then be analysed by standard statistical methods. The need to analyse very large and complex datasets has become a vital part of everyday life, particularly in the analysis of national problems in public health, environmental pollution, computer network sec ....Frontiers in Data Science: Analysing Distributions as Data. This project aims to develop the statistical foundations of a new approach to analysing large and complex data, based on building distributional approximations of the data, which can then be analysed by standard statistical methods. The need to analyse very large and complex datasets has become a vital part of everyday life, particularly in the analysis of national problems in public health, environmental pollution, computer network security and climate extremes. The project expects to change our way of thinking in how to be smarter about what data we use (and collect) for analysis, rather than relying on brute force analysis of large datasets. The project is expected to transform the knowledge base of the discipline, and the resulting techniques will enable across-the-board research advances for many industries and disciplines.Read moreRead less