The HIV/AIDS pandemic is currently the most challenging public health matter that faces third world countries, especially those in Sub-Saharan Africa. Ethiopia, in East Africa, with a generalised and highly heterogeneous epidemic, is no exception, with HIV/AIDS affecting most sectors of the economy. The first case of HIV in Ethiopia was reported in 1984. Since then, HIV/AIDS has become a major public health con cern, leading the Government of Ethiopia to declare a public health emergency in 2002. In 2011, the adult HIV/AIDS prevalence in Ethiopia was estimated at 1.5%. Approximately 1.2 million Ethiopians were living with HIV/AIDS in 2010. Surveys are an important and popular tool for collecting data. Analytical use of survey data especially health survey data has become very common, with a focus on the association of particular outcome variables with explanatory variables at the population level. In this study we used the data from the 2005 Ethiopian Demographic and Health Survey, (EDHS 2005), and identified key demographic, socioeconomic, sociocultural, behavioral and proximate determinants of HIV/AIDS risk factor. Usually most survey analysts ignore the complex survey design issues like clustering, stratification and unequal probability of selection (weights). This study deals with complex survey design and takes the design aspect into account, because failure to do so leads to bias parameters estimates and standard error, wide confidence intervals and statistical tests will be incorrect. In this study, three statistical approaches were used to analyse the complex survey data. The first approach was a survey logistic regression used to model the binary outcome (HIV serostatus) and set of explanatory variables (the dependence of the HIV risk factors). The difference between survey logistic regression and the ordinary logistic regression is that survey logistic regression approach takes the study design into account during analysis. The second approach was a multilevel logistic regression model, that assumed that the data structure in the population was hierarchical, and that individual within household was selected from clusters that were randomly selected from a national sampling frame. We considered a three-level model for our analysis. This second approach considered the results from Frequentist and a Bayesian multilevel models. Bayesian methods can provide accurate estimates of the parameters and the uncertainty associated with them. The third approach used was a Spatial models approach where model parameters were estimated under the Integrated Nested Laplace Approximation (INLA) paradigm. / Thesis (Ph.D.)-University of KwaZulu-Natal, Pietermaritzburg, 2013.

Catalyzed by protein kinases, phosphorylation is the most important post-translational modification in eukaryotes and is involved in the regulation of almost all cellular processes. Investigating phosphorylation events and how they change in response to different biological conditions is integral to understanding cellular signaling processes in general, as well as to defining the role of phosphorylation in health and disease. A recently-developed technology for studying phosphorylation events is the kinome microarray, which consists of several hundred "spots" arranged in a grid-like pattern on a glass slide. Each spot contains many peptides of a particular amino acid sequence chemically fixed to the slide, with different spots containing peptides with different sequences. Each peptide is a subsequence of a full protein, containing an amino acid residue that is known or suspected to undergo phosphorylation in vivo, as well as several surrounding residues. When a kinome microarray is exposed to cell lysate, the protein kinases in the lysate catalyze the phosphorylation of the peptides on the array. By measuring the degree to which the peptides comprising each spot are phosphorylated, insight can be gained into the upregulation or downregulation of signaling pathways in response to different biological treatments or conditions. There are two main computational challenges associated with kinome microarrays. The first is array design, which involves selecting the peptides to be included on a given array. The level of difficulty of this task depends largely on the number of phosphorylation sites that have been experimentally identified in the proteome of the organism being studied. For instance, thousands of phosphorylation sites are known for human and mouse, allowing considerable freedom to select peptides that are relevant to the problem being examined. In contrast, few sites are known for, say, honeybee and soybean. For such organisms, it is useful to expand the set of possible peptides by using computational techniques to predict probable phosphorylation sites. In this thesis, existing techniques for the computational prediction of phosphorylation sites are reviewed. In addition, two novel methods are described for predicting phosphorylation events in organisms with few known sites, with each method using a fundamentally different approach. The first technique, called PHOSFER, uses a random forest-based machine-learning strategy, while the second, called DAPPLE, takes advantage of sequence homology between known sites and the proteome of interest. Both methods are shown to allow quicker or more accurate predictions in organisms with few known sites than comparable previous techniques. Therefore, the use of kinome microarrays is no longer limited to the study of organisms having many known phosphorylation sites; rather, this technology can potentially be applied to any organism having a sequenced genome. It is shown that PHOSFER and DAPPLE are suitable for identifying phosphorylation sites in a wide variety of organisms, including cow, honeybee, and soybean. The second computational challenge is data analysis, which involves the normalization, clustering, statistical analysis, and visualization of data resulting from the arrays. While software designed for the analysis of DNA microarrays has also been used for kinome arrays, differences between the two technologies prompted the development of PIIKA, a software package specifically designed for the analysis of kinome microarray data. By comparing with methods used for DNA microarrays, it is shown that PIIKA improves the ability to identify biological pathways that are differentially regulated in a treatment condition compared to a control condition. Also described is an updated version, PIIKA 2, which contains improvements and new features in the areas of clustering, statistical analysis, and data visualization. Given the previous absence of dedicated tools for analyzing kinome microarray data, as well as their wealth of features, PIIKA and PIIKA 2 represent an important step in maximizing the scientific value of this technology. In addition to the above techniques, this thesis presents three studies involving biological applications of kinome microarray analysis. The first study demonstrates the existence of "kinotypes" - species- or individual-specific kinome profiles - which has implications for personalized medicine and for the use of model organisms in the study of human disease. The second study uses kinome analysis to characterize how the calf immune system responds to infection by the bacterium Mycobacterium avium subsp. paratuberculosis. Finally, the third study uses kinome arrays to study parasitism of honeybees by the mite Varroa destructor, which is thought to be a major cause of colony collapse disorder. In order to make the methods described above readily available, a website called the SAskatchewan PHosphorylation Internet REsource (SAPHIRE) has been developed. Located at the URL http://saphire.usask.ca, SAPHIRE allows researchers to easily make use of PHOSFER, DAPPLE, and PIIKA 2. These resources facilitate both the design and data analysis of kinome microarrays, making them an even more effective technique for studying cellular signaling.

Background: The overall quality of care has been defined in terms of a set of seven core attributes taken from contemporary conceptual frameworks for assessing primary healthcare systems. Attributes are assessed using sub-attribute questions picked from previously developed and validated national level survey instruments. Data has been collected through structured questionnaire survey utilizing Likert items and scale to capture respondents’ perceptions of care. Both descriptive and nonparametric statistical methods have been used for data analysis. Information on demographic factors helped to understand the response patterns across different cohort groups. Key objectives: 1) To determine the perception of patients and physicians regarding the overall quality of care and its constituent elements delivered through the primary healthcare clinics in Winnipeg. 2) To compare the perceptions about different quality of care attributes as expressed by participating patients and physicians. Results: Both patients and physicians have positive views about the overall quality of care (median score >=4 on a 1-6 scale). Regarding individual attributes, “Interpersonal communication” and “Respectfulness” received the highest average score (5) and long-term health management received the lowest score (2). Patient and physician responses were found to be statistically different for access, comprehensiveness and long-term health management. The long wait time for seeing a doctor appeared to be a widely shared concern – only 43% of the patients urgently needing to see a doctor could get a same-day appointment; for non-urgent cases, less than 3% got a same-day appointment. Patients with higher educational levels appeared to be more critical about the quality of care; conversely, patients in good health rated the quality of care attributes more favourably. Conclusion: Patients and physicians are generally satisfied with the overall quality of care. However, patients have identified issues related to access, comprehensiveness of care and long-term health management. Patients concerns were found to be consistent with national level results. Long wait time was also flagged as a key concern. Primary healthcare clinics should proactively seek patient feedback to identify issues and improve their quality of service.

Simulations based on a month of radar data from Florida, and a summer of radar data from Nelspruit, South Africa, were used to quantify the errors in the measurement of mean areal rainfall which arise simply as a result of the extreme variability of convective rainfall, even with perfect remote sensing instruments. The raingauge network measurement errors were established for random and regular network configurations using daily and monthly radar-rainfall accumulations over large areas. A relationship to predict the measurement error for mean areal rainfall using sparse networks as a function of raining area, number of gauges, and the variability of the rainfield was developed and tested. The manner in which the rainfield probability distribution is transformed under increasing spatial and temporal averaging was investigated from two perspectives. Firstly, an empirical relationship was developed to transform the probability distribution based on some measurement scale, into a distribution based on a standard measurement length. Secondly, a conceptual model based on multiplicative cascades was used to derive a scale independent probability distribution.

Least squares linear regression is a common tool in ecological research. One of the central assumptions of least squares linear regression is that the independent variable is measured without error. But this variable is measured with error whenever it is a sample mean. The significance of such contraventions is not regularly assessed in ecological studies. A simulation program was made to provide such an assessment. The program requires a hypothetical data set, and using estimates of S$ sp2$ it scatters the hypothetical data to simulate the effect of sampling error. A regression line is drawn through the scattered data, and SSE and r$ sp2$ are measured. This is repeated numerous times (e.g. 1000) to generate probability distributions for r$ sp2$ and SSE. From these distributions it is possible to assess the likelihood of the hypothetical data resulting in a given SSE or r$ sp2$. The method was applied to survey data used in a published TP-CHLa regression (Pace 1984). Beginning with a hypothetical, linear data set (r$ sp2$ = 1), simulated scatter due to sampling exceeded the SSE from the regression through the survey data about 30% of the time. Thus chances are 3 out of 10 that the level of uncertainty found in the surveyed TP-CHLa relationship would be observed if the true relationship were perfectly linear. If this is so, more precise and more comprehensive models will only be possible when better estimates of the means are available. This simulation approach should apply to all least squares regression studies that use sampled means, and should be especially relevant to studies that use log-transformed values.

This research focuses to develop some new techniques on statistical learning including methodology, computation and application. We also developed statistical quantification in nanomaterials. For a large number of random variables with temporal or spatial structures, we proposed shrink estimates of covariance matrix to account their Markov structures. The proposed method exploits the sparsity in the inverse covariance matrix in a systematic fashion. To deal with high dimensional data, we proposed a robust kernel principal component analysis for dimension reduction, which can extract the nonlinear structure of high dimension data more robustly. To build a prediction model more efficiently, we developed an active learning via sequential design to actively select the data points into the training set. By combining the stochastic approximation and D-optimal designs, the proposed method can build model with minimal time and effort. We also proposed factor logit-models with a large number of categories for classification. We show that the convergence rate of the classifier functions estimated from the proposed factor model does not rely on the number of categories, but only on the number of factors. It therefore can achieve better classification accuracy. For the statistical nano-quantification, a statistical approach is presented to quantify the elastic deformation of nanomaterials. We proposed a new statistical modeling technique, called sequential profile adjustment by regression (SPAR), to account for and eliminate the various experimental errors and artifacts. SPAR can automatically detect and remove the systematic errors and therefore gives more precise estimation of the elastic modulus.

Typescript. / Thesis (Ph. D.)--University of Hawaii at Manoa, 1984. / Bibliography: leaves 319-329. / Microfiche. / lMaster negative: Microfiche MS33170. / xxiii, 329 leaves, bound ill., maps 29 cm

This thesis develops a new method for predicting probabilities in a statistical parser so that more sophisticated probabilistic grammars can be used. A statistical parser uses a probabilistic grammar derived from a training corpus of hand-parsed sentences. The grammar is represented as a set of constructions - in a simple case these might be context-free rules. The probability of each construction in the grammar is then estimated by counting its relative frequency in the corpus. A crucial problem when building a probabilistic grammar is to select an appropriate level of granularity for describing the constructions being learned. The more constructions we include in our grammar, the more sophisticated a model of the language we produce. However, if too many different constructions are included, then our corpus is unlikely to contain reliable information about the relative frequency of many constructions. In existing statistical parsers two main approaches have been taken to choosing an appropriate granularity. In a non-lexicalised parser constructions are specified as structures involving particular parts-of-speech, thereby abstracting over individual words. Thus, in the training corpus two syntactic structures involving the same parts-of-speech but different words would be treated as two instances of the same event. In a lexicalised grammar the assumption is that the individual words in a sentence carry information about its syntactic analysis over and above what is carried by its part-of-speech tags. Lexicalised grammars have the potential to provide extremely detailed syntactic analyses; however, Zipf�s law makes it hard for such grammars to be learned. In this thesis, we propose a method for optimising the trade-off between informative and learnable constructions in statistical parsing. We implement a grammar which works at a level of granularity in between single words and parts-of-speech, by grouping words together using unsupervised clustering based on bigram statistics. We begin by implementing a statistical parser to serve as the basis for our experiments. The parser, based on that of Michael Collins (1999), contains a number of new features of general interest. We then implement a model of word clustering, which we believe is the first to deliver vector-based word representations for an arbitrarily large lexicon. Finally, we describe a series of experiments in which the statistical parser is trained using categories based on these word representations.

In this dissertation, application of multi-way analysis is extended into new areas of environmental chemistry, microbiology, electrochemistry and organometallic chemistry. Additionally new practical aspects of some of the multi-way analysis methods are discussed. Parallel Factor Analysis Two (PARAFAC2) is used to classify a wide range of weathered petroleum oils using GC-MS data. Various chemical and data analysis issues exist in the current methods of oil spill analysis are discussed and the proposed method is demonstrated to have potential to be employed in identification of source of oil spills. Two important practical aspects of PARAFAC2 are exploited to deal with chromatographic shifts and non-diagnostic peaks.GEneralized Multiplicative ANalysis Of VAriance (GEMANOVA) is applied to assess the bactericidal activity of new natural antibacterial extracts on three species of bacteria in different structure and oxidation forms and different concentrations. In this work while the applicability of traditional ANOVA is restricted due to the high interaction amongst the factors, GEMANOVA is shown to return robust and easily interpretable models which conform to the actual structure of the data. Peptide-modified electrochemical sensors are used to determine three metal cations of Cu2+, Cd2+ and Pb2+ simultaneously. Two sets of experiments are performed using a four-electrode system returning a three-way array of size (sample ?? current ?? electrode) and a single electrode resulting in a two-way data set of size (sample ?? current). The data of former is modeled by N-PLS and that latter using PLS. Despite the presence of highly overlapped voltammograms and several sources of non-linearity N-PLS returns reasonable models while PLS fails. An intramolecular hydroamination reaction is catalyzed by several organometallic catalysts to identify the most effective catalysts. The reaction of starting material in the presence of 72 different catalysts is monitored by UV-Vis at two time points, before and after heating the mixtures in an oven. PARAFAC is applied to the three-way data set of (sample ?? wavelength ?? time) to resolve the overlapped UV-Vis peaks and to identify the effective catalysts using the estimated relative concentration of product (loadings plot of the sample mode).

Ongoing globalisation has resulted in more liberalisation, integration, and competition among countries. An upshot of this has been higher levels of cross-border investment. Foreign direct investment (FDI), long considered an engine of growth, has led to widespread probe with its recent rapid spread. Nevertheless, while research on the contribution of FDI to host countries has concentrated heavily on the developed and developing economies, there has been a marked neglect of small, developed economies. This study proposes to focus on New Zealand, a country that falls within the latter category. The study seeks to verify econometrically the impact of FDI on the country through causality links with growth, trade, domestic investment and labour productivity. The analysis is based upon time-series data, the econometric techniques of single, autoregressive distributed lag (ARDL), and the multiple equations approach, vector error correction method (VECM). The study found that there have been substantial gains to the New Zealand economy. A positive effect of FDI on the variables mentioned above led to an improvement of the balance of payments through an increase in exports rather than in imports. Economic growth has mainly been achieved through FDI's impact on exports and domestic private investment. The dynamic innovation techniques indicated a bi-directional causality between FDI and the variables. The long-run causality, however, runs mainly from growth and labour productivity to FDI rather than in the opposite direction. Another noticeable feature is that New Zealand's regional agreement with Australia, Closer Economic Relations, has brought the country significant gains in terms of growth and development through FDI. Both the ARDL and VECM approaches suggest that for a small, developed country qualitative impacts are greater than quantitative ones. The policy implication is that maintaining sustainable economic growth with a positive domestic investment environment is vital for attracting foreign investors. New Zealand, while continuing to encourage inward FDI, should aim to channel it into 'innovative' tradable sectors. The challenge lies in providing the right kind of policy mix for this purpose.