Skip to main content


Seminar 05 June

Inferring genetic linkage maps from high-throughput sequencing data

Date: 05 June 2020, Friday

Time: 2pm

Speaker: Dr Matthew Schofield (University of Otago)

Abstract: Genetic maps are usually the starting point for many types of genetic analysis. They are one-dimensional representations of genetic inheritance across a chromosome. Genetic maps frequency are commonly inferred from estimates of a hidden Markov model (HMM) since only the expression and not the transmission of genetic information is observed. No general approaches exist for assessing the uncertainty of the map.

In this talk, we will obtain genetic maps and associated uncertainty for data arising from high-throughput sequencing (HTS). HTS technology provides high density data from a large numbers of individuals in a cost- and time-efficient manner. However, the observed data from HTS are more error prone than previous technologies. We first extend the HMM to account for error introduced by HTS. We then use a Bayesia…
Recent posts

Seminar 22 May 2020 3pm

The use of Fast Fourier Transforms and Generalized Poissonian distribution to study COVID Deaths

Date: 22 May 2020, Friday

Time: 3 pm

Speaker: A/Prof John Nichols (Texas A&M University)


The COVID Fatality data is often grouped into subsets that represent political boundaries, if these political boundaries represent unique compact urban areas fully contained in the urban sense then the application of the SEIR model appears to be somewhat applicable, but, if this is not the case, the assumptions that are made for the SEIR model may result in a poor predictive model. The use of Fast Fourier transforms of the residual data from an exponential regression analysis provides a method to estimate the frequency response of the residuals, which can be used to review the SEIR modelling of the urban area.  The second method is a GPD analysis of the daily ratio of the fatalities to the prior day, which may prove to be a method to determine unique compactness. Examples using NY and Texa…

Seminar 22 May 2020

Optimizing the Fitting of Linear Mixed Models - Comparing BLAS Subroutines in Isolation (no pun intended)

Date: 22 May 2020, Friday

Time: 2 pm

Speaker: Luke Mazur (Univeristy of Wollongong)


Linear mixed models arising from animal and plant breeding result in sparse sets of Mixed Model Equations with particular structures. An effective method of fitting these models is the Average Information (AI) algorithm, and the largest computational bottlenecks in the AI algorithm are the solution of these equations and the calculation of the Sparse Inverse Subset for the AI updating. There are a number of potential Basic Linear Algebra Sublibrary (BLAS) subroutines that can be used for these tasks, and potential candidates are investigated via a comparative experiment to see which combination of these is best.



Seminar 12 June 2020

Bayesian modelling of complex trajectories: a case study of covid-19

Date: 12 June 2020, Thursday

Time: 2 pm

Speaker: Prof. Kerrie Mengersen (Queensland University of Technology)

Abstract: Since the initial outbreak in Wuhan (Hubei, China) in December 2019, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the virus responsible for coronavirus disease 2019 (COVID-19), has rapidly spread to cause one of the most pressing challenges facing our world today: the COVID-19 pandemic. Within four months of the first reported cases, more than two and a half million cases were confirmed with over two hundred thousand deaths globally, and many countries had taken extreme measures to stop the spread. Although Bayesian models of epidemics are well known in the literature, modelling COVID-19 has been problematic because of the complexity of control responses that were implemented to contain the spread of the disease in different countries. In this presentation, I will describe…

Seminar 28 May 2020

Estimation of long-memory parameter in stationary and non-stationary curve time series

Date: 28 May 2020, Thursday

Time: 10 am

Speaker: A/Prof. Hanlin Shang (ANU)

Abstract: We study a functional version of fractionally integrated stationary and nonstationary time series, covering the functional unit root as a special case. The functional time series are projected onto a finite number of sub-spaces, the level of stationarity/non-stationary allowed to vary over them. Through the classic functional principal component analysis of the sample variance operator, we obtain the eigenvalues and eigenfunctions which span a sample version of the dominant subspace. Furthermore, we introduce a simple ratio criterion to consistently estimate the dimension of the dominant sub-space, and use a memory parameter estimator, such as local Whittle estimator, to estimate the memory parameter. Monte-Carlo simulation studies and empirical applications are given to examine the finite-sample performance of the…

Seminar 14 May 2020

Ensembles of Trees and CLT's: Inference and Machine Learning

Date: 14 May 2020, Thursday

Time: 10 am

Speaker: Professor Giles Hooker (ANU)

Abstract: This talk develops methods of statistical inference based around ensembles of decision trees: bagging, random forests, and boosting. Recent results have shown that when the bootstrap procedure in bagging methods is replaced by sub-sampling, predictions from these methods can be analyzed using the theory of U-statistics which have a limiting normal distribution. Moreover, the limiting variance that can be estimated within the sub-sampling structure.

Using this result, we can compare the predictions made by a model learned with a feature of interest, to those made by a model learned without it and ask whether the differences between these could have arisen by chance. By evaluating the model at a structured set of points we can also ask whether it differs significantly from an additive model. We demonstrate these results in an application t…

Seminar 08 May 2020 3pm

Application of statistical Quality control in clinical area

Date: 08 May, Friday

Time: 3-4pm

Speaker: A/Prof Mali Abdollahain (RMIT University)

Abstract: While statistical quality control and profile monitoring have been extensively used in manufacturing area, their application in clinical area has just been started. In clinical monitoring, there are always more than one quality characteristics of interest which are usually correlated. In such cases, multivariate control charts should be deployed to monitor the medical process. In manufacturing industry, Profile monitoring systems assist and help to identify factors related to an observed phenomenon, assess the effect of changing any factor/s on the event and predict the behaviour of the phenomenon under different situations. In many situations the quality and performance of a medical process may be better characterized and summarized by relationship between the response (dependent) variable and one or more explanatory (independent) v…

Seminar 08 May 2020

Forecasting nonlocal climate impacts for mobile marine species using extensions to empirical orthogonal function analysis

Date: 08 May 2020, Friday

Time: 2 pm

Speaker: Dr James Thorson (NOAA)

Abstract: Societal responses to COVID-19 have illustrated the great public value of accurate epidemiological forecasts; climate change has a similar potential to upend commerce and necessitates accurate decadal forecasts of community impacts. In this talk, I discuss modern extensions to Empirical Orthogonal Function (EOF) analysis and how it can be used to jointly analyze climate change and community ecology. EOF analysis is widely used to identify modes of variability from spatially distributed environmental measurements (e.g., the El NiƱo Southern Oscillation is primary mode of variability in sea surface temperatures in the Pacific Ocean), but is less common in community (or epidemiological) modelling to understand modes of community variability. I specifically argue that EOF analysis is one potent…

Seminar 01 May 2020 3pm

Depth for Curve Data and Applications

Date:Friday, 1 May 2020

Speaker: A/Prof Pierre Lafaye de Micheaux (UNSW)
Abstract:John W. Tukey (1975) defined statistical data depth as a function that determines the centrality of an arbitrary point with respect to a data cloud or to a probability measure. During the last decades, this seminal idea of data depth evolved into a powerful tool proving to be useful in various fields of science. Recently, extending the notion of data depth to the functional setting attracted a lot of attention among theoretical and applied statisticians. We go further and suggest a notion of data depth suitable for data represented as curves, or trajectories, which is independent of the parametrization. We show that our curve depth satisfies theoretical requirements of general depth functions that are meaningful for trajectories. We apply our methodology to diffusion tensor brain images and also to hurricane tracks.

Seminar 01 May 2020

Representativeness and Generalisability of Inference for Samples of Networks

Date: 01 May 2020, Friday

Time: 4 pm

Speaker: Dr Pavel Krivitsky (UNSW)

Abstract: Joint modelling of large samples of networks collected from similar settings—classrooms, households, etc.—has a long history, with a variety of methods available to pool information in model estimation and inference. In the exponential-family random graph modelling framework, these methods range from post-hoc two-stage meta-analyses to sophisticated multilevel approaches. However, relatively little attention has been devoted to the generalisability of this inference, especially when the sample of networks is effectively a convenience sample, and when the population of networks is heterogeneous in size and composition.

We consider two samples of within-household contact networks in Flanders, Belgium, which used very similar survey instruments but very different sampling designs: 1) a sample of 318 households, selected based on having …