User behaviour modelling in a multi-dimensional environment for personalization and recommendation

Handling information overload online, from the user's point of view is a big challenge, especially when the number of websites is growing rapidly due to growth in e-commerce and other related activities. Personalization based on user needs is the key to solving the problem of information overload. Personalization methods help in identifying relevant information, which may be liked by a user. User profile and object profile are the important elements of a personalization system. When creating user and object profiles, most of the existing methods adopt two-dimensional similarity methods based on vector or matrix models in order to find inter-user and inter-object similarity. Moreover, for recommending similar objects to users, personalization systems use the users-users, items-items and users-items similarity measures. In most cases similarity measures such as Euclidian, Manhattan, cosine and many others based on vector or matrix methods are used to find the similarities. Web logs are high-dimensional datasets, consisting of multiple users, multiple searches with many attributes to each. Two-dimensional data analysis methods may often overlook latent relationships that may exist between users and items. In contrast to other studies, this thesis utilises tensors, the high-dimensional data models, to build user and object profiles and to find the inter-relationships between users-users and users-items. To create an improved personalized Web system, this thesis proposes to build three types of profiles: individual user, group users and object profiles utilising decomposition factors of tensor data models. A hybrid recommendation approach utilising group profiles (forming the basis of a collaborative filtering method) and object profiles (forming the basis of a content-based method) in conjunction with individual user profiles (forming the basis of a model based approach) is proposed for making effective recommendations. A tensor-based clustering method is proposed that utilises the outcomes of popular tensor decomposition techniques such as PARAFAC, Tucker and HOSVD to group similar instances. An individual user profile, showing the user's highest interest, is represented by the top dimension values, extracted from the component matrix obtained after tensor decomposition. A group profile, showing similar users and their highest interest, is built by clustering similar users based on tensor decomposed values. A group profile is represented by the top association rules (containing various unique object combinations) that are derived from the searches made by the users of the cluster. An object profile is created to represent similar objects clustered on the basis of their similarity of features. Depending on the category of a user (known, anonymous or frequent visitor to the website), any of the profiles or their combinations is used for making personalized recommendations. A ranking algorithm is also proposed that utilizes the personalized information to order and rank the recommendations. The proposed methodology is evaluated on data collected from a real life car website. Empirical analysis confirms the effectiveness of recommendations made by the proposed approach over other collaborative filtering and content-based recommendation approaches based on two-dimensional data analysis methods.

Differential expression of chemokine and matrix re-modelling genes is associated with contrasting schistosome-induced hepatopathology in murine models

The pathological outcomes of schistosomiasis are largely dependent on the molecular and cellular mechanisms of the host immune response. In this study, we investigated the contribution of variations in host gene expression to the contrasting hepatic pathology observed between two inbred mouse strains following Schistosoma japonicum infection. Whole genome microarray analysis was employed in conjunction with histological and immunohistochemical analysis to define and compare the hepatic gene expression profiles and cellular composition associated with the hepatopathology observed in S. japonicum-infected BALB/c and CBA mice. We show that the transcriptional profiles differ significantly between the two mouse strains with high statistical confidence. We identified specific genes correlating with the more severe pathology associated with CBA mice, as well as genes which may confer the milder degree of pathology associated with BALB/c mice. In BALB/c mice, neutrophil genes exhibited striking increases in expression, which coincided with the significantly greater accumulation of neutrophils at granulomatous regions seen in histological sections of hepatic tissue. In contrast, up-regulated expression of the eosinophil chemokine CCL24 in CBA mice paralleled the cellular influx of eosinophils to the hepatic granulomas. Additionally, there was greater down-regulation of genes involved in metabolic processes in CBA mice, reflecting the more pronounced hepatic damage in these mice. Profibrotic genes showed similar levels of expression in both mouse strains, as did genes associated with Th1 and Th2 responses. However, imbalances in expression of matrix metalloproteinases (e.g. MMP12, MMP13) and tissue inhibitors of metalloproteinases (TIMP1) may contribute to the contrasting pathology observed in the two strains. Overall, these results provide a more complete picture of the molecular and cellular mechanisms which govern the pathological outcome of hepatic schistosomiasis. This improved understanding of the immunopathogenesis in the murine model schistosomiasis provides the basis for a better appreciation of the complexities associated with chronic human schistosomiasis.

Pedigree-free animal models : the relatedness matrix reloaded

Animal models typically require a known genetic pedigree to estimate quantitative genetic parameters. Here we test whether animal models can alternatively be based on estimates of relatedness derived entirely from molecular marker data. Our case study is the morphology of a wild bird population, for which we report estimates of the genetic variance-covariance matrices (G) of six morphological traits using three methods: the traditional animal model; a molecular marker-based approach to estimate heritability based on Ritland's pairwise regression method; and a new approach using a molecular genealogy arranged in a relatedness matrix (R) to replace the pedigree in an animal model. Using the traditional animal model, we found significant genetic variance for all six traits and positive genetic covariance among traits. The pairwise regression method did not return reliable estimates of quantitative genetic parameters in this population, with estimates of genetic variance and covariance typically being very small or negative. In contrast, we found mixed evidence for the use of the pedigree-free animal model. Similar to the pairwise regression method, the pedigree-free approach performed poorly when the full-rank R matrix based on the molecular genealogy was employed. However, performance improved substantially when we reduced the dimensionality of the R matrix in order to maximize the signal to noise ratio. Using reduced-rank R matrices generated estimates of genetic variance that were much closer to those from the traditional model. Nevertheless, this method was less reliable at estimating covariances, which were often estimated to be negative. Taken together, these results suggest that pedigree-free animal models can recover quantitative genetic information, although the signal remains relatively weak. It remains to be determined whether this problem can be overcome by the use of a more powerful battery of molecular markers and improved methods for reconstructing genealogies.

Accounting for management costs in sensitivity analyses of matrix population models

Traditional sensitivity and elasticity analyses of matrix population models have been used to inform management decisions, but they ignore the economic costs of manipulating vital rates. For example, the growth rate of a population is often most sensitive to changes in adult survival rate, but this does not mean that increasing that rate is the best option for managing the population because it may be much more expensive than other options. To explore how managers should optimize their manipulation of vital rates, we incorporated the cost of changing those rates into matrix population models. We derived analytic expressions for locations in parameter space where managers should shift between management of fecundity and survival, for the balance between fecundity and survival management at those boundaries, and for the allocation of management resources to sustain that optimal balance. For simple matrices, the optimal budget allocation can often be expressed as simple functions of vital rates and the relative costs of changing them. We applied our method to management of the Helmeted Honeyeater (Lichenostomus melanops cassidix; an endangered Australian bird) and the koala (Phascolarctos cinereus) as examples. Our method showed that cost-efficient management of the Helmeted Honeyeater should focus on increasing fecundity via nest protection, whereas optimal koala management should focus on manipulating both fecundity and survival simultaneously. These findings are contrary to the cost-negligent recommendations of elasticity analysis, which would suggest focusing on managing survival in both cases. A further investigation of Helmeted Honeyeater management options, based on an individual-based model incorporating density dependence, spatial structure, and environmental stochasticity, confirmed that fecundity management was the most cost-effective strategy. Our results demonstrate that decisions that ignore economic factors will reduce management efficiency. ©2006 Society for Conservation Biology.

Addressing issues in sparseness, ecological bias and formulation of the adjacency matrix in Bayesian spatio-temporal analysis of disease counts

The main objective of this PhD was to further develop Bayesian spatio-temporal models (specifically the Conditional Autoregressive (CAR) class of models), for the analysis of sparse disease outcomes such as birth defects. The motivation for the thesis arose from problems encountered when analyzing a large birth defect registry in New South Wales. The specific components and related research objectives of the thesis were developed from gaps in the literature on current formulations of the CAR model, and health service planning requirements. Data from a large probabilistically-linked database from 1990 to 2004, consisting of fields from two separate registries: the Birth Defect Registry (BDR) and Midwives Data Collection (MDC) were used in the analyses in this thesis. The main objective was split into smaller goals. The first goal was to determine how the specification of the neighbourhood weight matrix will affect the smoothing properties of the CAR model, and this is the focus of chapter 6. Secondly, I hoped to evaluate the usefulness of incorporating a zero-inflated Poisson (ZIP) component as well as a shared-component model in terms of modeling a sparse outcome, and this is carried out in chapter 7. The third goal was to identify optimal sampling and sample size schemes designed to select individual level data for a hybrid ecological spatial model, and this is done in chapter 8. Finally, I wanted to put together the earlier improvements to the CAR model, and along with demographic projections, provide forecasts for birth defects at the SLA level. Chapter 9 describes how this is done. For the first objective, I examined a series of neighbourhood weight matrices, and showed how smoothing the relative risk estimates according to similarity by an important covariate (i.e. maternal age) helped improve the model’s ability to recover the underlying risk, as compared to the traditional adjacency (specifically the Queen) method of applying weights. Next, to address the sparseness and excess zeros commonly encountered in the analysis of rare outcomes such as birth defects, I compared a few models, including an extension of the usual Poisson model to encompass excess zeros in the data. This was achieved via a mixture model, which also encompassed the shared component model to improve on the estimation of sparse counts through borrowing strength across a shared component (e.g. latent risk factor/s) with the referent outcome (caesarean section was used in this example). Using the Deviance Information Criteria (DIC), I showed how the proposed model performed better than the usual models, but only when both outcomes shared a strong spatial correlation. The next objective involved identifying the optimal sampling and sample size strategy for incorporating individual-level data with areal covariates in a hybrid study design. I performed extensive simulation studies, evaluating thirteen different sampling schemes along with variations in sample size. This was done in the context of an ecological regression model that incorporated spatial correlation in the outcomes, as well as accommodating both individual and areal measures of covariates. Using the Average Mean Squared Error (AMSE), I showed how a simple random sample of 20% of the SLAs, followed by selecting all cases in the SLAs chosen, along with an equal number of controls, provided the lowest AMSE. The final objective involved combining the improved spatio-temporal CAR model with population (i.e. women) forecasts, to provide 30-year annual estimates of birth defects at the Statistical Local Area (SLA) level in New South Wales, Australia. The projections were illustrated using sixteen different SLAs, representing the various areal measures of socio-economic status and remoteness. A sensitivity analysis of the assumptions used in the projection was also undertaken. By the end of the thesis, I will show how challenges in the spatial analysis of rare diseases such as birth defects can be addressed, by specifically formulating the neighbourhood weight matrix to smooth according to a key covariate (i.e. maternal age), incorporating a ZIP component to model excess zeros in outcomes and borrowing strength from a referent outcome (i.e. caesarean counts). An efficient strategy to sample individual-level data and sample size considerations for rare disease will also be presented. Finally, projections in birth defect categories at the SLA level will be made.

Bayesian hybrid algorithms and models : implementation and associated issues

This thesis addresses computational challenges arising from Bayesian analysis of complex real-world problems. Many of the models and algorithms designed for such analysis are ‘hybrid’ in nature, in that they are a composition of components for which their individual properties may be easily described but the performance of the model or algorithm as a whole is less well understood. The aim of this research project is to after a better understanding of the performance of hybrid models and algorithms. The goal of this thesis is to analyse the computational aspects of hybrid models and hybrid algorithms in the Bayesian context. The first objective of the research focuses on computational aspects of hybrid models, notably a continuous finite mixture of t-distributions. In the mixture model, an inference of interest is the number of components, as this may relate to both the quality of model fit to data and the computational workload. The analysis of t-mixtures using Markov chain Monte Carlo (MCMC) is described and the model is compared to the Normal case based on the goodness of fit. Through simulation studies, it is demonstrated that the t-mixture model can be more flexible and more parsimonious in terms of number of components, particularly for skewed and heavytailed data. The study also reveals important computational issues associated with the use of t-mixtures, which have not been adequately considered in the literature. The second objective of the research focuses on computational aspects of hybrid algorithms for Bayesian analysis. Two approaches will be considered: a formal comparison of the performance of a range of hybrid algorithms and a theoretical investigation of the performance of one of these algorithms in high dimensions. For the first approach, the delayed rejection algorithm, the pinball sampler, the Metropolis adjusted Langevin algorithm, and the hybrid version of the population Monte Carlo (PMC) algorithm are selected as a set of examples of hybrid algorithms. Statistical literature shows how statistical efficiency is often the only criteria for an efficient algorithm. In this thesis the algorithms are also considered and compared from a more practical perspective. This extends to the study of how individual algorithms contribute to the overall efficiency of hybrid algorithms, and highlights weaknesses that may be introduced by the combination process of these components in a single algorithm. The second approach to considering computational aspects of hybrid algorithms involves an investigation of the performance of the PMC in high dimensions. It is well known that as a model becomes more complex, computation may become increasingly difficult in real time. In particular the importance sampling based algorithms, including the PMC, are known to be unstable in high dimensions. This thesis examines the PMC algorithm in a simplified setting, a single step of the general sampling, and explores a fundamental problem that occurs in applying importance sampling to a high-dimensional problem. The precision of the computed estimate from the simplified setting is measured by the asymptotic variance of the estimate under conditions on the importance function. Additionally, the exponential growth of the asymptotic variance with the dimension is demonstrated and we illustrates that the optimal covariance matrix for the importance function can be estimated in a special case.

A review of simulation models for railway systems

With the advances in computer hardware and software development techniques in the past 25 years, digital computer simulation of train movement and traction systems has been widely adopted as a standard computer-aided engineering tool [1] during the design and development stages of existing and new railway systems. Simulators of different approaches and scales are used extensively to investigate various kinds of system studies. Simulation is now proven to be the cheapest means to carry out performance predication and system behaviour characterisation. When computers were first used to study railway systems, they were mainly employed to perform repetitive but time-consuming computational tasks, such as matrix manipulations for power network solution and exhaustive searches for optimal braking trajectories. With only simple high-level programming languages available at the time, full advantage of the computing hardware could not be taken. Hence, structured simulations of the whole railway system were not very common. Most applications focused on isolated parts of the railway system. It is more appropriate to regard those applications as primarily mechanised calculations rather than simulations. However, a railway system consists of a number of subsystems, such as train movement, power supply and traction drives, which inevitably contains many complexities and diversities. These subsystems interact frequently with each other while the trains are moving; and they have their special features in different railway systems. To further complicate the simulation requirements, constraints like track geometry, speed restrictions and friction have to be considered, not to mention possible non-linearities and uncertainties in the system. In order to provide a comprehensive and accurate account of system behaviour through simulation, a large amount of data has to be organised systematically to ensure easy access and efficient representation; the interactions and relationships among the subsystems should be defined explicitly. These requirements call for sophisticated and effective simulation models for each component of the system. The software development techniques available nowadays allow the evolution of such simulation models. Not only can the applicability of the simulators be largely enhanced by advanced software design, maintainability and modularity for easy understanding and further development, and portability for various hardware platforms are also encouraged. The objective of this paper is to review the development of a number of approaches to simulation models. Attention is, in particular, given to models for train movement, power supply systems and traction drives. These models have been successfully used to enable various ‘what-if’ issues to be resolved effectively in a wide range of applications, such as speed profiles, energy consumption, run times etc.

Development of zonal cartilage constructs : effects of chondrocyte subpopulation, compressive stimulation, and culture models

Articular cartilage is a highly resilient tissue located at the ends of long bones. It has a zonal structure, which has functional significance in load-bearing. Cartilage does not spontaneously heal itself when damaged, and untreated cartilage lesions or age-related wear often lead to osteoarthritis (OA). OA is a degenerative condition that is highly prevalent, age-associated, and significantly affects patient mobility and quality of life. There is no cure for OA, and patients usually resort to replacing the biological joint with an artificial prosthesis. An alternative approach is to dynamically regenerate damaged or diseased cartilage through cartilage tissue engineering, where cells, materials, and stimuli are combined to form new cartilage. However, despite extensive research, major limitations remain that have prevented the wide-spread application of tissue-engineered cartilage. Critically, there is a dearth of information on whether autologous chondrocytes obtained from OA patients can be used to successfully generate cartilage tissues with structural hierarchy typically found in normal articular cartilage. I aim to address these limitations in this thesis by showing that chondrocyte subpopulations isolated from macroscopically normal areas of the cartilage can be used to engineer stratified cartilage tissues and that compressive loading plays an important role in zone-dependent biosynthesis of these chondrocytes. I first demonstrate that chondrocyte subpopulations from the superficial (S) and middle/deep (MD) zones of OA cartilage are responsive to compressive stimulation in vitro, and that the effect of compression on construct quality is zone-dependent. I also show that compressive stimulation can influence pericelluar matrix production, matrix metalloproteinase secretion, and cytokine expression in zonal chondrocytes in an alginate hydrogel model. Subsequently, I focus on recreating the zonal structure by forming layered constructs using the alginate-released chondrocyte (ARC) method either with or without polymeric scaffolds. Resulting zonal ARC constructs had hyaline morphology, and expressed cartilage matrix molecules such as proteoglycans and collagen type II in both scaffold-free and scaffold-based approaches. Overall, my findings demonstrate that chondrocyte subpopulations obtained from OA joints respond sensitively to compressive stimulation, and are able to form cartilaginous constructs with stratified organization similar to native cartilage using the scaffold-free and scaffold-based ARC technique. The ultimate goal in tissue engineering is to help provide improved treatment options for patients suffering from debilitating conditions such as OA. Further investigations in developing functional cartilage replacement tissues using autologous chondrocytes will bring us a step closer to improving the quality of life for millions of OA patients worldwide.

Eyes on 3D-Current 3D biomimetic disease concept models and potential applications in age-related macular degeneration

Three dimensional cellular models that mimic disease are being increasingly investigated and have opened an exciting new research area into understanding pathomechanisms. The advantage of 3D in vitro disease models is that they allow systematic and in-depth studies of physiological and pathophysiological processes with less costs and ethical concerns that have arisen with animal models. The purpose of the 3D approach is to allow crosstalk between cells and microenvironment, and with cues from the microenvironment, cells can assemble their niche similar to in vivo conditions. The use of 3D models for mimicking disease processes such as cancer, osteoarthritis etc., is only emerging and allows multidisciplinary teams consisting of tissue engineers, biologist biomaterial scientists and clinicians to work closely together. While in vitro systems require rigorous testing before they can be considered as replicates of the in vivo model, major steps have been made, suggesting that they will become powerful tools for studying physiological and pathophysiological processes. This paper aims to summarize some of the existing 3D models and proposes a novel 3D model of the eye structures that are involved in the most common cause of blindness in the Western World, namely age-related macular degeneration (AMD).

Sources of ultrafine particles and chemical species along a traffic corridor : comparison of the results from two receptor models

Particulate matter is common in our environment and has been linked to human health problems particularly in the ultrafine size range. A range of chemical species have been associated with particulate matter and of special concern are the hazardous chemicals that can accentuate health problems. If the sources of such particles can be identified then strategies can be developed for the reduction of air pollution and consequently, the improvement of the quality of life. In this investigation, particle number size distribution data and the concentrations of chemical species were obtained at two sites in Brisbane, Australia. Source apportionment was used to determine the sources (or factors) responsible for the particle size distribution data. The apportionment was performed by Positive Matrix Factorisation (PMF) and Principal Component Analysis/Absolute Principal Component Scores (PCA/APCS), and the results were compared with information from the gaseous chemical composition analysis. Although PCA/APCS resolved more sources, the results of the PMF analysis appear to be more reliable. Six common sources identified by both methods include: traffic 1, traffic 2, local traffic, biomass burning, and two unassigned factors. Thus motor vehicle related activities had the most impact on the data with the average contribution from nearly all sources to the measured concentrations higher during peak traffic hours and weekdays. Further analyses incorporated the meteorological measurements into the PMF results to determine the direction of the sources relative to the measurement sites, and this indicated that traffic on the nearby road and intersection was responsible for most of the factors. The described methodology which utilised a combination of three types of data related to particulate matter to determine the sources could assist future development of particle emission control and reduction strategies.

Matrix metalloproteinases 2 and 9 (gelatinases A and B) expression in malignant mesothelioma and benign pleura

Matrix metalloproteinases (MMPs), in particular the gelatinases (MMP-2 and -9), play a significant role in tumour invasion and angiogenesis. The expression and activities of MMPs have not been characterised in malignant mesothelioma (MM) tumour samples. In a prospective study, gelatinase activity was evaluated in homogenised supernatants of snap frozen MM (n = 35), inflamed pleura (IP, n = 12) and uninflammed pleura (UP, n = 14) tissue specimens by semiquantitative gelatin zymography. Matrix metalloproteinases were correlated with clinicopathological factors and with survival using Kaplan-Meier and Cox proportional hazard models. In MM, pro- and active MMP-2 levels were significantly greater than for MMP-9 (P = 0.006, P<0.001). Active MMP-2 was significantly greater in MM than in UP (P=0.04). MMP-2 activity was equivalent between IP and MM, but both pro- and active MMP-9 activities were greater in IP (P=0.02, P=0.009). While there were trends towards poor survival with increasing total and pro-MMP-2 activity (P=0.08) in univariate analysis, they were both independent poor prognostic factors in multivariate analysis in conjunction with weight loss (pro-MMP-2 P = 0.03, total MMP-2 P = 0.04). Total and pro-MMP-2 also contributed to the Cancer and Leukemia Group B prognostic groups. MMP-9 activities were not prognostic. Matrix metalloproteinases, and in particular MMP-2, the most abundant gelatinase, may play an important role in MM tumour growth and metastasis. Agents that reduce MMP synthesis and/or activity may have a role to play in the management of MM. © 2003 Cancer Research UK.

Local stress-strain distribution and load transfer across cartilage matrix at micro-scale using combined microscopy-based finite element method

Articular cartilage is the load-bearing tissue that consists of proteoglycan macromolecules entrapped between collagen fibrils in a three-dimensional architecture. To date, the drudgery of searching for mathematical models to represent the biomechanics of such a system continues without providing a fitting description of its functional response to load at micro-scale level. We believe that the major complication arose when cartilage was first envisaged as a multiphasic model with distinguishable components and that quantifying those and searching for the laws that govern their interaction is inadequate. To the thesis of this paper, cartilage as a bulk is as much continuum as is the response of its components to the external stimuli. For this reason, we framed the fundamental question as to what would be the mechano-structural functionality of such a system in the total absence of one of its key constituents-proteoglycans. To answer this, hydrated normal and proteoglycan depleted samples were tested under confined compression while finite element models were reproduced, for the first time, based on the structural microarchitecture of the cross-sectional profile of the matrices. These micro-porous in silico models served as virtual transducers to produce an internal noninvasive probing mechanism beyond experimental capabilities to render the matrices micromechanics and several others properties like permeability, orientation etc. The results demonstrated that load transfer was closely related to the microarchitecture of the hyperelastic models that represent solid skeleton stress and fluid response based on the state of the collagen network with and without the swollen proteoglycans. In other words, the stress gradient during deformation was a function of the structural pattern of the network and acted in concert with the position-dependent compositional state of the matrix. This reveals that the interaction between indistinguishable components in real cartilage is superimposed by its microarchitectural state which directly influences macromechanical behavior.

Selecting volatility forecasting models for portfolio allocation purposes

Techniques for evaluating and selecting multivariate volatility forecasts are not yet understood as well as their univariate counterparts. This paper considers the ability of different loss functions to discriminate between a set of competing forecasting models which are subsequently applied in a portfolio allocation context. It is found that a likelihood-based loss function outperforms its competitors, including those based on the given portfolio application. This result indicates that considering the particular application of forecasts is not necessarily the most effective basis on which to select models.

Kinematic models for manoeuvring and seakeeping of marine vessels

The motion of marine vessels has traditionally been studied using two different approaches: manoeuvring and seakeeping. These two approaches use different reference frames and coordinate systems to describe the motion. This paper derives the kinematic models that characterize the transformation of motion variables (position, velocity, accelerations) and forces between the different coordinate systems used in these theories. The derivations hereby presented are done in terms of the formalism adopted in robotics. The advantage of this formulation is the use of matrix notation and operations. As an application, the transformation of linear equations of motion used in seakeeping into body-fixed coordinates is considered for both zero and forward speed.