Mixture models for high-dimensional clustering with applications to tumour classification, network intrusion, and text classification. This project will benefit the Australian Society as a whole by developing statistical methodology for the clustering of high-dimensional data. In particular, it will develop a novel and efficient model for extracting useful information from subpopulations. It thus has wide applicability to improving the quality and validity of applied research in most industries ....Mixture models for high-dimensional clustering with applications to tumour classification, network intrusion, and text classification. This project will benefit the Australian Society as a whole by developing statistical methodology for the clustering of high-dimensional data. In particular, it will develop a novel and efficient model for extracting useful information from subpopulations. It thus has wide applicability to improving the quality and validity of applied research in most industries in Australia. More specifically, it is to be applied here to classify brain tumours and detect network intruders. This cross-disciplinary project will contribute to Australia's economic of public health, protect Australia from crime, and strength Australian researchers' capacity and capability of participating in this emerging science.Read moreRead less
Bayesian Statistical Inference for Implicitly defined Probability Models. Bayesian statistics has recently been used to provide solutions for a large number of hitherto intractable problems in science and technology. The success of Bayesian statistics has mainly been due to the application of so-called Markov chain Monte Carlo computational techniques. We aim to improve these algorithms, by providing fast, simple and efficient computational implementations. We will use the results to give ins ....Bayesian Statistical Inference for Implicitly defined Probability Models. Bayesian statistics has recently been used to provide solutions for a large number of hitherto intractable problems in science and technology. The success of Bayesian statistics has mainly been due to the application of so-called Markov chain Monte Carlo computational techniques. We aim to improve these algorithms, by providing fast, simple and efficient computational implementations. We will use the results to give insight by carefully quantifying and modelling uncertainty for such topics as the transmission rate of infectious diseases, the spatial distribution of plant and animal species, investigating biological theory for the genome of a virus, and changes in human fertility.Read moreRead less
Motor Unit Numbers Estimation (MUNE) using Bayesian statistical methodology for monitoring of progression of neuromuscular diseases. A means of objectively measuring the pathology of a neuromuscular disease involving motor unit loss, such as motor neuron disease, is much needed. This will be achieved by using newly developed electrophysiological techniques and developing new Bayesian statistical methodology to determine the number of motor units that supply a muscle. Our innovations will reliabl ....Motor Unit Numbers Estimation (MUNE) using Bayesian statistical methodology for monitoring of progression of neuromuscular diseases. A means of objectively measuring the pathology of a neuromuscular disease involving motor unit loss, such as motor neuron disease, is much needed. This will be achieved by using newly developed electrophysiological techniques and developing new Bayesian statistical methodology to determine the number of motor units that supply a muscle. Our innovations will reliably determine the number of motor units that supply a muscle in both normal subjects and in diseased patients with loss of motor nerves. This will enable the monitoring of disease progression. An outcome will be a software package that can be used with standard electrophysiology machines.Read moreRead less
Statistical methods for analysing maps in the visual brain. This project aims to apply Gaussian process methods, a Bayesian approach for data analysis, to analyse data from brain imaging experiments. Discovering the principles of functional brain architecture requires analysing data from functional imaging technologies. However, these technologies produce very noisy data which is difficult to interpret. This project will apply Gaussian process methods to study data from optical imaging and funct ....Statistical methods for analysing maps in the visual brain. This project aims to apply Gaussian process methods, a Bayesian approach for data analysis, to analyse data from brain imaging experiments. Discovering the principles of functional brain architecture requires analysing data from functional imaging technologies. However, these technologies produce very noisy data which is difficult to interpret. This project will apply Gaussian process methods to study data from optical imaging and functional magnetic resonance imaging of the visual brain. This is expected to reveal critical information about how normal brain structure changes with development and sensory experience. The statistical methods developed should be applicable within and beyond neuroscience, and may ultimately help improve the diagnosis of human health disorders.Read moreRead less
Applications of Bayesian methods in Genomics and Comparative Genomics. Bayesian statistics provides a unified and versatile approach to problems of data analysis, inference and hypothesis testing. This project will involve the application of Bayesian methods to four topics of commercial and scientific importance in the fields of Genomics and Comparative Genomics. The four topics are: data analysis for a novel DNA sequencing technology, investigating genomic structure using multiple change-point ....Applications of Bayesian methods in Genomics and Comparative Genomics. Bayesian statistics provides a unified and versatile approach to problems of data analysis, inference and hypothesis testing. This project will involve the application of Bayesian methods to four topics of commercial and scientific importance in the fields of Genomics and Comparative Genomics. The four topics are: data analysis for a novel DNA sequencing technology, investigating genomic structure using multiple change-point analysis, phlogenetic inference with multiple genes and detection of incongruent phylogenies. The overall goal of the project is to advance understanding of the structure, function and evolution of genomes.Read moreRead less
Nonlinear Econometric Modelling: A Complex Systems Perspective. It is becoming increasingly accepted that economic systems are both complex and adaptive. However, this introduces a range of problems in constructing, estimating and testing economic models using time series data. In this project, this problem will be addressed through the formulation and implementation of a new methodology and associated techniques. These techniques will allow a researcher to use information obtained from a set o ....Nonlinear Econometric Modelling: A Complex Systems Perspective. It is becoming increasingly accepted that economic systems are both complex and adaptive. However, this introduces a range of problems in constructing, estimating and testing economic models using time series data. In this project, this problem will be addressed through the formulation and implementation of a new methodology and associated techniques. These techniques will allow a researcher to use information obtained from a set of nonlinearity tests to determine which type of nonlinear model provides the best representation of a data generating mechanism. Selected high frequency financial and macroeconomic data (for the US and Australia) will be used in the study. This research is intended to change the direction and emphasis of econometric modelling and promises to have a fundamental impact on forecasting and policy evaluation methods.
Read moreRead less
Unsupervised learning of finite mixture models in data mining applications. The extraction of useful information from massively large databases is known as data mining. Its broad but vague goal is to find "interesting structure" in the data, which typically leads to breaking the data into clusters. To this end, we consider the fast, efficient, and automatic learning of finite mixture models in hugh data sets without any prior knowledge of the structure. This probabilistic approach to the discove ....Unsupervised learning of finite mixture models in data mining applications. The extraction of useful information from massively large databases is known as data mining. Its broad but vague goal is to find "interesting structure" in the data, which typically leads to breaking the data into clusters. To this end, we consider the fast, efficient, and automatic learning of finite mixture models in hugh data sets without any prior knowledge of the structure. This probabilistic approach to the discovery and validation of group structure in data mining applications will considerably enhance knowledge management and decision support in science, industry, and government.
Read moreRead less
On-line and Incremental EM-based Neural Networks: Application to Hospital Utlilization and Gene Expression Data. Artificial neural networks have been widely applied as universal classifiers in many fields, such as biomedicine. However, misunderstanding of fundamental statistical principles, which can cause misleading findings, has been frequently observed in the literature. This project aims to integrate statistical methodologies in neural networks to provide a unified approach to improve its ....On-line and Incremental EM-based Neural Networks: Application to Hospital Utlilization and Gene Expression Data. Artificial neural networks have been widely applied as universal classifiers in many fields, such as biomedicine. However, misunderstanding of fundamental statistical principles, which can cause misleading findings, has been frequently observed in the literature. This project aims to integrate statistical methodologies in neural networks to provide a unified approach to improve its applicability and efficiency in implementation. The system developed from this proposed cross-disciplinary research will be applied to hospital utilization data (hospital morbidity database, Western Australia) and gene expression data (DNA microarrays databases, Harvard University). This collaborative research will advance the international standard of Australian research communities.
Read moreRead less
Discovery Early Career Researcher Award - Grant ID: DE130100819
Funder
Australian Research Council
Funding Amount
$281,600.00
Summary
Measuring the improbable: optimal Monte Carlo methods for rare event simulation of maxima of dependent random variables. Some events occurring with low frequency can have dramatic consequences: natural catastrophes, economic crises, system malfunctions. Estimating their probabilities is a very difficult problem. This project will develop new simulation methods capable of delivering the most precise and efficient estimators for the probabilities of such events.
Linkage Infrastructure, Equipment And Facilities - Grant ID: LE0453501
Funder
Australian Research Council
Funding Amount
$406,097.00
Summary
A Computational Research Grid Serving Regional and Metropolitan Queensland. This project will advance scientific discovery through the development of an integrated, user-friendly computational grid. It significantly enhances Queensland's research capability and infrastructure by delivering state-of-the-art computational resources to researchers at the collaborating institutions and other Queensland and Australia researchers. New supercomputer systems will be integrated into a Queensland wide com ....A Computational Research Grid Serving Regional and Metropolitan Queensland. This project will advance scientific discovery through the development of an integrated, user-friendly computational grid. It significantly enhances Queensland's research capability and infrastructure by delivering state-of-the-art computational resources to researchers at the collaborating institutions and other Queensland and Australia researchers. New supercomputer systems will be integrated into a Queensland wide computational grid being developed by the Queensland Parallel Supercomputing Foundation - an initiative supported by the Queensland State Government. New grid technologies will be employed so that the highest level of support is provided to researchers. This ensures that the facility is used effectively, allowing high-quality research to be efficiently conducted.Read moreRead less