Hierarchical models for 2D presence/absence data having ambiguous zeroes: With a biogeographical case study on dingo behaviour

This dissertation is primarily an applied statistical modelling investigation, motivated by a case study comprising real data and real questions. Theoretical questions on modelling and computation of normalization constants arose from pursuit of these data analytic questions. The essence of the thesis can be described as follows. Consider binary data observed on a two-dimensional lattice. A common problem with such data is the ambiguity of zeroes recorded. These may represent zero response given some threshold (presence) or that the threshold has not been triggered (absence). Suppose that the researcher wishes to estimate the effects of covariates on the binary responses, whilst taking into account underlying spatial variation, which is itself of some interest. This situation arises in many contexts and the dingo, cypress and toad case studies described in the motivation chapter are examples of this. Two main approaches to modelling and inference are investigated in this thesis. The first is frequentist and based on generalized linear models, with spatial variation modelled by using a block structure or by smoothing the residuals spatially. The EM algorithm can be used to obtain point estimates, coupled with bootstrapping or asymptotic MLE estimates for standard errors. The second approach is Bayesian and based on a three- or four-tier hierarchical model, comprising a logistic regression with covariates for the data layer, a binary Markov Random field (MRF) for the underlying spatial process, and suitable priors for parameters in these main models. The three-parameter autologistic model is a particular MRF of interest. Markov chain Monte Carlo (MCMC) methods comprising hybrid Metropolis/Gibbs samplers is suitable for computation in this situation. Model performance can be gauged by MCMC diagnostics. Model choice can be assessed by incorporating another tier in the modelling hierarchy. This requires evaluation of a normalization constant, a notoriously difficult problem. Difficulty with estimating the normalization constant for the MRF can be overcome by using a path integral approach, although this is a highly computationally intensive method. Different methods of estimating ratios of normalization constants (N Cs) are investigated, including importance sampling Monte Carlo (ISMC), dependent Monte Carlo based on MCMC simulations (MCMC), and reverse logistic regression (RLR). I develop an idea present though not fully developed in the literature, and propose the Integrated mean canonical statistic (IMCS) method for estimating log NC ratios for binary MRFs. The IMCS method falls within the framework of the newly identified path sampling methods of Gelman & Meng (1998) and outperforms ISMC, MCMC and RLR. It also does not rely on simplifying assumptions, such as ignoring spatio-temporal dependence in the process. A thorough investigation is made of the application of IMCS to the three-parameter Autologistic model. This work introduces background computations required for the full implementation of the four-tier model in Chapter 7. Two different extensions of the three-tier model to a four-tier version are investigated. The first extension incorporates temporal dependence in the underlying spatio-temporal process. The second extensions allows the successes and failures in the data layer to depend on time. The MCMC computational method is extended to incorporate the extra layer. A major contribution of the thesis is the development of a fully Bayesian approach to inference for these hierarchical models for the first time. Note: The author of this thesis has agreed to make it open access but invites people downloading the thesis to send her an email via the 'Contact Author' function.

Experimental evaluation of hotspot identification methods

Identifying crash “hotspots”, “blackspots”, “sites with promise”, or “high risk” locations is standard practice in departments of transportation throughout the US. The literature is replete with the development and discussion of statistical methods for hotspot identification (HSID). Theoretical derivations and empirical studies have been used to weigh the benefits of various HSID methods; however, a small number of studies have used controlled experiments to systematically assess various methods. Using experimentally derived simulated data—which are argued to be superior to empirical data, three hot spot identification methods observed in practice are evaluated: simple ranking, confidence interval, and Empirical Bayes. Using simulated data, sites with promise are known a priori, in contrast to empirical data where high risk sites are not known for certain. To conduct the evaluation, properties of observed crash data are used to generate simulated crash frequency distributions at hypothetical sites. A variety of factors is manipulated to simulate a host of ‘real world’ conditions. Various levels of confidence are explored, and false positives (identifying a safe site as high risk) and false negatives (identifying a high risk site as safe) are compared across methods. Finally, the effects of crash history duration in the three HSID approaches are assessed. The results illustrate that the Empirical Bayes technique significantly outperforms ranking and confidence interval techniques (with certain caveats). As found by others, false positives and negatives are inversely related. Three years of crash history appears, in general, to provide an appropriate crash history duration.

Hierarchical Bayesian models for estimating the extent of plant pest invasions

Plant biosecurity requires statistical tools to interpret field surveillance data in order to manage pest incursions that threaten crop production and trade. Ultimately, management decisions need to be based on the probability that an area is infested or free of a pest. Current informal approaches to delimiting pest extent rely upon expert ecological interpretation of presence / absence data over space and time. Hierarchical Bayesian models provide a cohesive statistical framework that can formally integrate the available information on both pest ecology and data. The overarching method involves constructing an observation model for the surveillance data, conditional on the hidden extent of the pest and uncertain detection sensitivity. The extent of the pest is then modelled as a dynamic invasion process that includes uncertainty in ecological parameters. Modelling approaches to assimilate this information are explored through case studies on spiralling whitefly, Aleurodicus dispersus and red banded mango caterpillar, Deanolis sublimbalis. Markov chain Monte Carlo simulation is used to estimate the probable extent of pests, given the observation and process model conditioned by surveillance data. Statistical methods, based on time-to-event models, are developed to apply hierarchical Bayesian models to early detection programs and to demonstrate area freedom from pests. The value of early detection surveillance programs is demonstrated through an application to interpret surveillance data for exotic plant pests with uncertain spread rates. The model suggests that typical early detection programs provide a moderate reduction in the probability of an area being infested but a dramatic reduction in the expected area of incursions at a given time. Estimates of spiralling whitefly extent are examined at local, district and state-wide scales. The local model estimates the rate of natural spread and the influence of host architecture, host suitability and inspector efficiency. These parameter estimates can support the development of robust surveillance programs. Hierarchical Bayesian models for the human-mediated spread of spiralling whitefly are developed for the colonisation of discrete cells connected by a modified gravity model. By estimating dispersal parameters, the model can be used to predict the extent of the pest over time. An extended model predicts the climate restricted distribution of the pest in Queensland. These novel human-mediated movement models are well suited to demonstrating area freedom at coarse spatio-temporal scales. At finer scales, and in the presence of ecological complexity, exploratory models are developed to investigate the capacity for surveillance information to estimate the extent of red banded mango caterpillar. It is apparent that excessive uncertainty about observation and ecological parameters can impose limits on inference at the scales required for effective management of response programs. The thesis contributes novel statistical approaches to estimating the extent of pests and develops applications to assist decision-making across a range of plant biosecurity surveillance activities. Hierarchical Bayesian modelling is demonstrated as both a useful analytical tool for estimating pest extent and a natural investigative paradigm for developing and focussing biosecurity programs.

QoC Elastic Scheduling for Real-Time Control Systems

The elastic task model, a significant development in scheduling of real-time control tasks, provides a mechanism for flexible workload management in uncertain environments. It tells how to adjust the control periods to fulfill the workload constraints. However, it is not directly linked to the quality-of-control (QoC) management, the ultimate goal of a control system. As a result, it does not tell how to make the best use of the system resources to maximize the QoC improvement. To fill in this gap, a new feedback scheduling framework, which we refer to as QoC elastic scheduling, is developed in this paper for real-time process control systems. It addresses the QoC directly through embedding both the QoC management and workload adaptation into a constrained optimization problem. The resulting solution for period adjustment is in a closed-form expressed in QoC measurements, enabling closed-loop feedback of the QoC to the task scheduler. Whenever the QoC elastic scheduler is activated, it improves the QoC the most while still meeting the system constraints. Examples are given to demonstrate the effectiveness of the QoC elastic scheduling.

Evaluation of financial incentives as a quality improvement strategy in the public hospital context : clinician attitudes, design variables, and economic costs

In 2008, a three-year pilot ‘pay for performance’ (P4P) program, known as ‘Clinical Practice Improvement Payment’ (CPIP) was introduced into Queensland Health (QHealth). QHealth is a large public health sector provider of acute, community, and public health services in Queensland, Australia. The organisation has recently embarked on a significant reform agenda including a review of existing funding arrangements (Duckett et al., 2008). Partly in response to this reform agenda, a casemix funding model has been implemented to reconnect health care funding with outcomes. CPIP was conceptualised as a performance-based scheme that rewarded quality with financial incentives. This is the first time such a scheme has been implemented into the public health sector in Australia with a focus on rewarding quality, and it is unique in that it has a large state-wide focus and includes 15 Districts. CPIP initially targeted five acute and community clinical areas including Mental Health, Discharge Medication, Emergency Department, Chronic Obstructive Pulmonary Disease, and Stroke. The CPIP scheme was designed around key concepts including the identification of clinical indicators that met the set criteria of: high disease burden, a well defined single diagnostic group or intervention, significant variations in clinical outcomes and/or practices, a good evidence, and clinician control and support (Ward, Daniels, Walker & Duckett, 2007). This evaluative research targeted Phase One of implementation of the CPIP scheme from January 2008 to March 2009. A formative evaluation utilising a mixed methodology and complementarity analysis was undertaken. The research involved three research questions and aimed to determine the knowledge, understanding, and attitudes of clinicians; identify improvements to the design, administration, and monitoring of CPIP; and determine the financial and economic costs of the scheme. Three key studies were undertaken to ascertain responses to the key research questions. Firstly, a survey of clinicians was undertaken to examine levels of knowledge and understanding and their attitudes to the scheme. Secondly, the study sought to apply Statistical Process Control (SPC) to the process indicators to assess if this enhanced the scheme and a third study examined a simple economic cost analysis. The CPIP Survey of clinicians elicited 192 clinician respondents. Over 70% of these respondents were supportive of the continuation of the CPIP scheme. This finding was also supported by the results of a quantitative altitude survey that identified positive attitudes in 6 of the 7 domains-including impact, awareness and understanding and clinical relevance, all being scored positive across the combined respondent group. SPC as a trending tool may play an important role in the early identification of indicator weakness for the CPIP scheme. This evaluative research study supports a previously identified need in the literature for a phased introduction of Pay for Performance (P4P) type programs. It further highlights the value of undertaking a formal risk assessment of clinician, management, and systemic levels of literacy and competency with measurement and monitoring of quality prior to a phased implementation. This phasing can then be guided by a P4P Design Variable Matrix which provides a selection of program design options such as indicator target and payment mechanisms. It became evident that a clear process is required to standardise how clinical indicators evolve over time and direct movement towards more rigorous ‘pay for performance’ targets and the development of an optimal funding model. Use of this matrix will enable the scheme to mature and build the literacy and competency of clinicians and the organisation as implementation progresses. Furthermore, the research identified that CPIP created a spotlight on clinical indicators and incentive payments of over five million from a potential ten million was secured across the five clinical areas in the first 15 months of the scheme. This indicates that quality was rewarded in the new QHealth funding model, and despite issues being identified with the payment mechanism, funding was distributed. The economic model used identified a relative low cost of reporting (under $8,000) as opposed to funds secured of over $300,000 for mental health as an example. Movement to a full cost effectiveness study of CPIP is supported. Overall the introduction of the CPIP scheme into QHealth has been a positive and effective strategy for engaging clinicians in quality and has been the catalyst for the identification and monitoring of valuable clinical process indicators. This research has highlighted that clinicians are supportive of the scheme in general; however, there are some significant risks that include the functioning of the CPIP payment mechanism. Given clinician support for the use of a pay–for-performance methodology in QHealth, the CPIP scheme has the potential to be a powerful addition to a multi-faceted suite of quality improvement initiatives within QHealth.

Bayesian methodology for genetics of complex diseases

Genetic research of complex diseases is a challenging, but exciting, area of research. The early development of the research was limited, however, until the completion of the Human Genome and HapMap projects, along with the reduction in the cost of genotyping, which paves the way for understanding the genetic composition of complex diseases. In this thesis, we focus on the statistical methods for two aspects of genetic research: phenotype definition for diseases with complex etiology and methods for identifying potentially associated Single Nucleotide Polymorphisms (SNPs) and SNP-SNP interactions. With regard to phenotype definition for diseases with complex etiology, we firstly investigated the effects of different statistical phenotyping approaches on the subsequent analysis. In light of the findings, and the difficulties in validating the estimated phenotype, we proposed two different methods for reconciling phenotypes of different models using Bayesian model averaging as a coherent mechanism for accounting for model uncertainty. In the second part of the thesis, the focus is turned to the methods for identifying associated SNPs and SNP interactions. We review the use of Bayesian logistic regression with variable selection for SNP identification and extended the model for detecting the interaction effects for population based case-control studies. In this part of study, we also develop a machine learning algorithm to cope with the large scale data analysis, namely modified Logic Regression with Genetic Program (MLR-GEP), which is then compared with the Bayesian model, Random Forests and other variants of logic regression.

The impact of knowledge integration on enterprise system success

Despite promising benefits and advantages, there are reports of failures and low realisation of benefits in Enterprise System (ES) initiatives. Among the research on the factors that influence ES success, there is a dearth of studies on the knowledge implications of multiple end-user groups using the same ES application. An ES facilitates the work of several user groups, ranging from strategic management, management, to operational staff, all using the same system for multiple objectives. Given the fundamental characteristics of ES – integration of modules, business process views, and aspects of information transparency – it is necessary that all frequent end-users share a reasonable amount of common knowledge and integrate their knowledge to yield new knowledge. Recent literature on ES implementation highlights the importance of Knowledge Integration (KI) for implementation success. Unfortunately, the importance of KI is often overlooked and little about the role of KI in ES success is known. Many organisations do not achieve the potential benefits from their ES investment because they do not consider the need or their ability to integrate their employees’ knowledge. This study is designed to improve our understanding of the influence of KI among ES end-users on operational ES success. The three objectives of the study are: (I) to identify and validate the antecedents of KI effectiveness, (II) to investigate the impact of KI effectiveness on the goodness of individuals’ ES-knowledge base, and (III) to examine the impact of the goodness of individuals’ ES-knowledge base on the operational ES success. For this purpose, we employ the KI factors identified by Grant (1996) and an IS-impact measurement model from the work of Gable et al. (2008) to examine ES success. The study derives its findings from data gathered from six Malaysian companies in order to obtain the three-fold goal of this thesis as outlined above. The relationships between the antecedents of KI effectiveness and its consequences are tested using 188 responses to a survey representing the views of management and operational employment cohorts. Using statistical methods, we confirm three antecedents of KI effectiveness and the consequences of the antecedents on ES success are validated. The findings demonstrate a statistically positive impact of KI effectiveness of ES success, with KI effectiveness contributing to almost one-third of ES success. This research makes a number of contributions to the understanding of the influence of KI on ES success. First, based on the empirical work using a complete nomological net model, the role of KI effectiveness on ES success is evidenced. Second, the model provides a theoretical lens for a more comprehensive understanding of the impact of KI on the level of ES success. Third, restructuring the dimensions of the knowledge-based theory to fit the context of ES extends its applicability and generalisability to contemporary Information Systems. Fourth, the study develops and validates measures for the antecedents of KI effectiveness. Fifth, the study demonstrates the statistically significant positive influence of the goodness of KI on ES success. From a practical viewpoint, this study emphasises the importance of KI effectiveness as a direct antecedent of ES success. Practical lessons can be drawn from the work done in this study to empirically identify the critical factors among the antecedents of KI effectiveness that should be given attention.