Search

© 2020, International Association for Mathematical Geosciences. Some methods for analysing mineral prospectivity, especially the weights of evidence technique, require the predictor variables to be binary values. When the original evidence data are numerical values, such as geochemical indices, they can be converted to binary values by thresholding. When the evidence layer is a spatial feature such as a geological fault system, it can be converted to a binary predictor by buffering at a suitable cut-off distance. This paper reviews methods for selecting the best threshold or cut-off value and compares their performance. The review covers techniques which are well known in prospectivity analysis as well as unfamiliar techniques borrowed from other literature. Methods include maximisation of the estimated contrast, Studentised contrast, χ2 test statistic, Youden criterion, statistical likelihood, Akman–Raftery criterion, and curvature of the capture–efficiency curve. We identify connections between the different methods, and we highlight a common technical error in their application. Simulation experiments indicate that the Youden criterion has the best performance for selection of the threshold or cut-off value, assuming that a simple binary threshold relationship truly holds. If the relationship between predictor and prospectivity is more complicated, then the likelihood method is the most easily adaptable. The weights-of-evidence contrast performs poorly overall. These conclusions are supported by our analysis of data from the Murchison goldfields, Western Australia. We also propose a bootstrap method for calculating standard errors and confidence intervals for the location of the threshold.

Using a structured fusion penalty in regression models containing only categorical explanatory variables yields patterns of indicator variables that are simpler and more easily interpretable than regressions produced using the more commonly-used LASSO and group LASSO penalties. We construct logistic regression models for predicting stillbirth from categorical explanatory variables and demonstrate that using a structured fusion penalty produces regressions that are easier to interpret yet yield similar predictive ability.

The recent initiative by the Actuaries Institute to incorporate a data science/analytics unit at the Honours or Masters level, as well as the changes to the undergraduate curriculum that will start in 2020, means that actuarial science educators need to consider embedding data analysis and analytics throughout the undergraduate curriculum. At Curtin, we have been trialling ways of doing just that in a coherent fashion, from first-year to third-year units, so that students see data analysis as an integral part of becoming a practicing actuary. We were motivated by: 1. The current funding reality, which dictates that units in one degree course serve as units in other courses. At Curtin, for example, we have introduced two new majors: Data Science, and Applied Statistics, both of which have data analytic and computational components, and so many units common to these courses have to do double- and even triple-duty; 2. The fact that many of our actuarial students are finding work as data analysts, and could become even more competitive in the marketplace were they to learn additional skills in computing and data analysis; and 3. The need to modernize the actuarial science curriculum, even before the changes that will take place in 2020. In this talk, I will outline some of the ways in which we have modified tuition and assessment patterns in several units to incorporate computing, data analysis, notions of reproducible analyses, computer-based assessments, and project work. In addition, I will point out what has worked well, as well as some of the barriers we have faced in integrating components of data analytic and computational thinking and practice throughout the entire actuarial science curriculum.

A common existing approach to modeling rainfall extremes employs a spatial Bayesian hierarchical model, where latent Gaussian processes are specified on distributional parameters in order to pool spatial information. The data are taken to be conditionally independent given the latent processes, and so spatial dependence arises only through this conditional structure. This methodology can be extended by incorporating an explicit max-stable dependence structure, which therefore produces more realistic spatial surfaces, and removes the assumption of conditional independence. We further extend the max-stable methodology to incorporate the joint modeling of rainfall data at different accumulation durations. We therefore pool information across both space and accumulation duration within a broad framework which includes an explicit specification of spatial dependence. Our model can be used to derive inferences at ungauged sites, and easily incorporates missing values. Our methodology is applied to a dataset of pluviometer records recorded at 182 meteorological stations located on the Central Coast of New South Wales, Australia. For each station, rainfall data are accumulated over 16 different durations ranging from 5 min to 7 days. The model is fitted using Markov chain Monte Carlo simulation, employing auxiliary variables so that exact Bayesian inference can be performed. We present estimated parameters and posterior inferences for isohyetal maps and intensity-duration-frequency curves at selected sites of interest, and compare our inferences with those derived from the standard latent variable model.

This paper describes a statistical modelling framework for the characterisation of rainfall extremes over a region of interest. Using a Bayesian hierarchical approach, the data are assumed to follow the generalised extreme value distribution, whose parameters are modelled as spatial Gaussian processes in the latent process layer. We also integrate a parametric relationship between precipitation maxima accumulated over increasing durations. The inference of the model parameters is thus improved by pooling information across both space and accumulation duration. In addition, we propose and investigate two different approaches for the integration of daily and sub-daily rainfall data within the framework. We also demonstrate how information from a regional climate model can be integrated to enable the investigation of future projections of extreme rainfall characteristics. We apply the proposed methodology to precipitation datasets from two large-scale study regions located on the east coast of Australia. The models are fitted using Markov chain Monte Carlo simulations, and we present estimated model parameters and posterior inferences of return levels at various durations and sites of interest. We demonstrate the effectiveness of the framework in spatially extrapolating the inference to locations other than those at which direct rainfall measurements are available. We also provide comparisons between rainfall extremes at various durations obtained for the current climate and those based on future projections from a regional climate model. Both methods proposed for the integration of daily and sub-daily records were found to yield similar results in terms of model performance and computational requirements.

As part of a special issue on natural hazards, this paper reviews the current state of scientific knowledge of Australian heatwaves. Over recent years, progress has been made in understanding both the causes of and changes to heatwaves. Relationships between atmospheric heatwaves and large-scale and synoptic variability have been identified, with increasing trends in heatwave intensity, frequency and duration projected to continue throughout the 21st century. However, more research is required to further our understanding of the dynamical interactions of atmospheric heatwaves, particularly with the land surface. Research into marine heatwaves is still in its infancy, with little known about driving mechanisms, and observed and future changes. In order to address these knowledge gaps, recommendations include: focusing on a comprehensive assessment of atmospheric heatwave dynamics; understanding links with droughts; working towards a unified measurement framework; and investigating observed and future trends in marine heatwaves. Such work requires comprehensive and long-term collaboration activities. However, benefits will extend to the international community, thus addressing global grand challenges surrounding these extreme events.

The new edition of Australian Rainfall and Runoff (hereafter designated ARR 2015) contains completely revised design rainfall Intensity-Frequency-Duration (IFD) curves prepared by the Bureau of Meteorology. These curves are estimated from current climate observations. The Fifth Assessment Report of the Intergovernmental Panel on Climate Change states that "extreme precipitation events over most of the mid-latitude land masses and over wet tropical regions will very likely become more intense and more frequent by the end of this century, as global mean surface temperature increases." This suggests that the current-climate IFD curves may become unsuitable for infrastructure design in future decades. While ARR 2015 includes an interim guideline on incorporating climate change into design flood estimation, the guidance is based on a 'broad brush' approach due to the paucity of published regionally specific results. As a first step towards bridging this gap, a pilot project titled 'Rainfall Intensity-Frequency-Duration (IFD) Relationships under Climate Change' was commissioned in June 2013 and administered by Engineers Australia. Its principal aim was to provide insight into how the new design rainfall IFD curves might be affected by anthropogenic climate change. This paper describes the background for and components of the investigation, the challenges involved, the major research findings and recommendations for further action.

The Intergovernmental Panel on Climate Change (IPCC, 2007) has predicted an increase in extreme rainfall due to climate change, which may also lead to an increase in natural hazards such as flooding. These hazards can result in damage to infrastructure and agriculture, and may even result in injury or loss of life. Consequently, there is a need for accurate analysis and projection of extreme rainfall and its potential impacts. For example, understanding the relationship between rainfall intensity, frequency, and duration is important for the design and safety of infrastructure so that it can withstand extreme rainfall events. This relationship is described graphically by intensity-frequency-duration (IFD) curves. Estimating IFD curves and their associated uncertainty as accurately as possible is critical as it may help reduce the human and economic impacts that result from extreme rainfall events. In this paper, we examine two methods for modelling extreme rainfall spatially: regional frequency analysis (RFA) and a Bayesian hierarchical model (BHM). We produce IFD estimates from both methods and compare the results. We find that for some locations, the RFA and BHM estimates are similar, and for other locations, they are different. We discuss the importance of uncertainty estimates and demonstrate the flexibility of the BHM for producing such measures of uncertainty.

Colorectal cancer (CRC) is the second most common cause of cancer-related death worldwide and places a major economic burden on the global health care system. The time frame for development from premalignant to malignant disease typically spans 10-15 years, and this latent period provides an ideal opportunity for early detection and intervention to improve patient outcomes. Currently, early diagnosis of CRC is hampered by a lack of suitable non-invasive biomarkers that are clinically or economically acceptable for population-based screening. New blood-based protein biomarkers for early detection of CRC are therefore urgently required. The success of clinical biomarker discovery and validation studies is critically dependent on understanding and adjusting for potential experimental, analytical, and biological factors that can interfere with the robust interpretation of results. In this review we outline some important considerations for research groups undertaking biomarker research with exemplars from our studies. Implementation of experimental strategies to minimise the potential effects of these problems will facilitate the identification of panels of biomarkers with the sensitivity and specificity required for the development of successful tests for the early detection and surveillance of CRC.

Climate and weather variables such as rainfall, temperature, and pressure are indicators for hazards such as tropical cyclones, floods, and fires. The impact of these events can be due to a single variable being in an extreme state, but more often it is the result of a combination of variables not all of which are necessarily extreme. Here, the combination of variables or events that lead to an extreme impact is referred to as a compound event. Any given compound event will depend upon the nature and number of physical variables, the range of spatial and temporal scales, the strength of dependence between processes, and the perspective of the stakeholder who defines the impact. Modeling compound events is a large, complex, and interdisciplinary undertaking. To facilitate this task we propose the use of influence diagrams for defining, mapping, analyzing, modeling, and communicating the risk of the compound event. Ultimately, a greater appreciation of compound events will lead to further insight and a changed perspective on how impact risks are associated with climate-related hazards.

Understanding weather and climate extremes is important for assessing, and adapting to, the potential impacts of climate change. The design of hydraulic structures such as dams, drainage and sewers, for instance, relies in part on accurate information regarding patterns of extreme rainfall occurring at various locations and at different durations. Deriving this kind of information is challenging from a statistical viewpoint because a lot of information must be extracted from very little data. In this paper, we describe the use of a spatial Bayesian hierarchical model (BHM) for characterising rainfall extremes over a region of interest, using historical records of precipitation data from a network of rainfall stations. The rainfall extremes are assumed to have a generalised extreme value (GEV) distribution, with the shape, scale and location parameters representing the underlying variables of the BHM’s process layer. These parameters are modelled as a linear regression over spatial covariates (latitude and longitude) with additive spatially-correlated random process. This spatial process leads to more precise estimates of rainfall extremes at gauged locations, and also allows the inference of parameters at ungauged locations. Furthermore, it also mitigates the limitations imposed by short rainfall records in that it allows the model to “borrow strength” from neighbouring sites, thereby reducing the uncertainty at both gauged and ungauged locations. Making use of r-largest order statistics in the data layer further allows the integration of multiple yearly rainfall amounts instead of the annual maximum only.The proposed BHM uses a parametric representation that links the GEV scale parameter obtained for different accumulation durations. This approach leads to two additional process parameters, and allows the use of pluviometer data accumulated over a range of durations, thereby also increasing the amount of data available for inference. A main advantage of the Bayesia

The Minimum Noise Fraction (MNF) transform is widely used in the remote sensing and image processing communities, because it is usually better than the Principal Components (PC) transform at compressing and ordering multispectral and hyperspectral images in terms of image “quality”. The MNF transform is also invariant to invertible (i.e. non-singular) linear transformations of multispectral/hyperspectral data, a property not shared by the PC transform. This general invariance property of the MNF transform is proved. Three examples of the general invariance property are provided and discussed: (i) invariance to scaling, (ii) invariance to certain types of background correction, and (iii) invariance to different types of noise.

In many statistical downscaling methods, atmospheric variables are chosen by using a combination of expert knowledge with empirical measures such as correlations and partial correlations. In this short communication, we describe the use of a fast, sparse variable selection method, known as RaVE, for selecting atmospheric predictors, and illustrate its use on rainfall occurrence at stations in South Australia. We show that RaVE generates parsimonious models that are both sensible and interpretable, and whose results compare favourably to those obtained by a non-homogeneous hidden Markov model (Hughes et al., 1999). © 2011.

Summary: NetRaVE is a small suite of R functions for generating dependency networks using sparse regression methods. Such networks provide an alternative to interpreting 'top n lists' of genes arising out of an analysis of microarray data, and they provide a means of organizing and visualizing the resulting information in a manner that may suggest relationships between genes. Availability: NetRaVE is freely available for academic use and has been tested in R 2.10.1 under Windows XP, Linux and Mac OS X. Contact: harri.kiiveri@csiro.au. Supplementary information: Supplementary data are available at Bioinformatics online. © The Author 2010. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org.

The iterated constrained endmembers (ICE) algorithm is a new method of unmixing hyperspectral images that combines aspects of multivariate curve resolution (MCR) methods in chemometrics and unmixing algorithms in remote sensing. Like many MCR methods, ICE also estimates pure components, or endmembers, via alternating least squares; however, it is explicitly based on a convex geometry model and estimation is carried out in a subspace of reduced dimensionality defined by the minimum noise fraction (MNF) transform. In this paper, we describe the ICE algorithm and its properties. We also illustrate its use on a hyperspectral image of cervical tissue. The unmixing of hyperspectral images presents some unique challenges, and we also outline where further development is required. Copyright © 2008 John Wiley & Sons, Ltd.