Infant and child mortality rates are one of the health indicators in a given community or country. It is the fourth millennium development goal that by 2015, all the united nations member countries are expected to have reduced their infant and child mortality rates by two-thirds. Uganda is one of those countries in sub-Saharan Africa with high infant and child mortality rates and therefore has the need to find out the factors strongly associated to these high rates in order to provide alternative or maintain the existing interventions. The Uganda Demographic Health Survey (UDHS) funded by USAID, UNFPA, UNICEF, Irish Aid and the United kingdom government provides a data set which is rich in information. This information has attracted many researchers and some of it can be used to help Uganda monitor her infant and child mortality rates to achieve the fourth millennium goal. Survival analysis techniques and frailty modelling is a well developed statistical tool in analysing time to event data. These methods were adopted in this thesis to examine factors affecting under-five child mortality in Uganda using the UDHS data for 2011 using R and STATA software. Results obtained by fitting the Cox-proportional hazard model and frailty models and drawing inference using both the Frequentists and Bayesian approach showed that, Demographic factors (sex of the household head, sex of the child and number of births in the past one year) are strongly associated with high under-five child mortality rates. Heterogeneity or unobserved covariates were found to be signifcant at household but insignifcant at community level. / Thesis (M.Sc.)-University of KwaZulu-Natal, Pietermaritzburg, 2013.

Survival analysis is a well developed area which explores time to single event analysis. In some cases, however, such methods may not adequately capture the disease process as the disease progression may involve intermediate events of interest. Multistate models incorporate multiple events or states. This thesis proposes to demystify the theory of multistate models through an application based approach. We present the key components of multistate models, relevant derivations, model diagnostics and techniques for modeling the effect of covariates on transition intensities. The methods that are developed in the thesis are applied to HIV and TB data partly sourced from CAPRISA and the HPP programmes in the University of KwaZulu-Natal. HIV progression is investigated through the application of a five state Markov model with reversible transitions such that state 1: CD4 count 500, state 2: 350 CD4 count < 500, state 3: 200 CD4 count < 350, state 4: CD4 count < 200 and state 5: ARV initiation. The mean sojourn time in each state and transition probabilities are presented as well as the effect of covariates namely age, gender and baseline CD4 count on transition rates. A key finding, consistent with previous research, is that the rate of decline in CD4 count tends to decrease at lower levels of the marker. Further, patients enrolling with a CD4 count less than 350 had a far lower chance of immune recovery and a substantially higher chance of immune deterioration compared to patients with a higher CD4 count. We noted that older patients tend to progress more rapidly through the disease than younger patients. / Thesis (M.Sc.)-University of KwaZulu-Natal, Westville, 2011.

No abstract available. / Thesis (M.Sc.)-University of Durban-Westville, 1992.

The option pricing theory set forth by Black and Scholes assumes that the underlying asset can be modeled by Geometric Brownian motion, with the Brownian motion being the driving force of uncertainty. Recent empirical studies, Dotsis, Psychoyios & Skiadopolous (2007) [17], suggest that the use of Brownian motion alone is insufficient in accurately describing the evolution of the underlying asset. A more realistic description of the underlying asset’s dynamics would be to include random jumps in addition to that of the Brownian motion. The concept of including jumps in the asset price model leads us naturally to the concept of a L'evy process. L'evy processes serve as a building block for stochastic processes that include jumps in addition to Brownian motion. In this dissertation we first examine the structure and nature of an arbitrary L'evy process. We then introduce the stochastic integral for L'evy processes as well as the extended version of Itˆo’s lemma, we then identify exponential L'evy processes that can serve as Radon-Nikod'ym derivatives in defining new probability measures. Equipped with our knowledge of L'evy processes we then implement this process in a financial context with the L'evy process serving as driving source of uncertainty in some stock price model. In particular we look at jump-diffusion models such as Merton’s(1976) [37] jump-diffusion model and the jump-diffusion model proposed by Kou and Wang (2004) [30]. As the L'evy processes we consider have more than one source of randomness we are faced with the difficulty of pricing options in an incomplete market. The options that we shall consider shall be mainly European in nature, where exercise can only occur at maturity. In addition to the vanilla calls and puts we independently derive a closed form solution for an exchange option under Merton’s jump-diffusion model making use of conditioning arguments and stochastic integral representations. We also examine some exotic options under the Kou and Wang model such as barrier options and lookback options where the solution to the option price is derived in terms of Laplace transforms. We then develop the Kou and Wang model to include only positive jumps, under this revised model we compute the value of a perpetual put option along with the optimal exercise point. Keywords Derivative pricing, L'evy processes, exchange options, stochastic integration. / Thesis (M.Sc.)-University of KwaZulu-Natal, Westville, 2009.

Dissertation (PhD)--University of Stellenbosch, 2002 / ENGLISH ABSTRACT: Singular spectrum analysis (SSA) originated in the field of Physics. The technique is non-parametric by nature and inter alia finds application in atmospheric sciences, signal processing and recently in financial markets. The technique can handle a very broad class of time series that can contain combinations of complex periodicities, polynomial or exponential trend. Forecasting techniques are reviewed in this study, and a new coordinate free joint-horizon k-period-ahead forecasting formulation is derived. The study also considers model selection in SSA, from which it become apparent that forward validation results in more stable model selection. The roots of SSA are outlined and distributional assumptions of signal senes are considered ab initio. Pitfalls that arise in the multivariate statistical theory are identified. Different approaches of recurrent one-period-ahead forecasting are then reviewed. The forecasting approaches are all supplied in algorithmic form to ensure effortless adaptation to computer programs. Theoretical considerations, underlying the forecasting algorithms, are also considered. A new coordinate free joint-horizon kperiod- ahead forecasting formulation is derived and also adapted for the multichannel SSA case. Different model selection techniques are then considered. The use of scree-diagrams, phase space portraits, percentage variation explained by eigenvectors, cross and forward validation are considered in detail. The non-parametric nature of SSA essentially results in the use of non-parametric model selection techniques. Finally, the study also considers a commercial software package that is available and compares it with Fortran code, which was developed as part of the study. / AFRIKAANSE OPSOMMING: Singulier spektraalanalise (SSA) het sy oorsprong in die Fisika. Die tegniek is nieparametries van aard en vind toepassing in velde soos atmosferiese wetenskappe, seinprossesering en onlangs in finansiële markte. Die tegniek kan 'n wye verskeidenheid tydreekse hanteer wat kombinasies van komplekse periodisiteite, polinomiese- en eksponensiële tendense insluit. Vooruitskattingstegnieke word ook in hierdie studie beskou, en 'n nuwe koërdinaatvrye gesamentlike horison k-periodevooruitskattingformulering word afgelei. Die studie beskou ook model seleksie in SSA, waaruit duidelik blyk dat voorwaartse validasie meer stabiele model seleksie tot gevolg het. Die agtergrond van SSA word ab initio geskets en verdelingsaannames van seinreekse beskou. Probleemgevalle wat voorkom in die meervoudige statistiese teorie word duidelik geïdentifiseer. Verskeie tegnieke van herhalende toepassing van een-periode-vooruitskatting word daarna beskou. Die benaderings tot vooruitskatting word in algororitmiese formaat verskaf wat die aanpassing na rekenaarprogrammering vergemaklik. Teoretiese vraagstukke, onderliggend aan die vooruitskattings-algortimes, word ook beskou. 'n Nuwe koërdinaatvrye gesamentlike horison k-periode-vooruitskattingsformulering word afgelei en aangepas vir die multikanaal SSA geval. Verskillende model seleksie tegnieke is ook beskou. Die gebruik van "scree"- diagramme, fase ruimte diagramme, persentasie variasie verklaar deur eievektore, kruis- en voorwaartse validasie word ook aangespreek. Die nie-parametriese aard van SSA noop die gebruik van nie-parametriese model seleksie tegnieke. Die studie vergelyk laastens 'n kommersiële sagtewarepakket met die Fortran bronkode wat as deel van hierdie studie ontwikkel is.

Eucalypt trees are one of tree species used for the manufacturing of papers in South Africa. The manufacturing of paper consists of cooking the wood with chemicals until obtaining a pulp. The wood is made of different cells. The shape and structure of these cells, called wood anatomical characteristics are important for the quality of paper. In addition, the anatomical characteristics of wood are influenced by environmental factors like climatic factors, soil compositions etc…. In this study we investigated the effects of the climatic factors (temperature, rainfall, solar radiation, relative humidity, and wind speed) on wood anatomical characteristics of two Eucalyptus clones, a GC (Eucalyptus grandis × camuldulensis) and a GU (Eucalyptus grandis × urophylla). Nine trees per clone have been selected. Two sets of data have been collected for this study. The first set of data was eleven anatomical characteristics of the wood formed daily over a period of five years. The second set of data was the daily measurement of temperature, rainfall, solar radiation, relative humidity and wind speed in the experimental area. Wood is made of two kinds of cell, the fibres and the vessels. The fibres are used for the strength and support of the tree and the vessels for the nutrition. Eleven characteristics related to those cells have been measured (diameter, wall thickness, frequency). These characteristics are highly correlated. To reduce the number of response variables, the principal component analysis was used and the first four principal components accounts for about 95% of the total variation. Based on the weights associated with each component the first four principal components were labelled as vessel dimension (VD), fibre dimension (FD), fibre wall (FW) and vessel frequency (VF). The longitudinal linear mixed model with age, season, temperature, rainfall, solar radiation, relative humidity and wind speed as the fixed effects factors and tree as random effect factor was fitted to the data. From time series modelling result, lagged order of climatic variables were identified and these lagged climatic variables were included in the model. To account for the physical characteristic of the trees we included the effect of diameter at breast height, stem radius, daily radial increment, and the suppression or dominance of the tree in the model. It was found that wood anatomical characteristics of the two clones were more affected by climatic variables when the tree was on juvenile stage as compared to mature stage. / Thesis (M.Sc.)-University of KwaZulu-Natal, Pietermaritzburg, 2010.

The most commonly used applications of hidden-layer feed forward neural networks are to fit curves to regression data or to provide a surface from which a classification rule can be found. From a statistical viewpoint, the principle underpinning these networks is that of nonparametric regression with sigmoidal curves being located and scaled so that their sum approximates the data well, and the underlying mechanism is that of nonlinear regression, with the weights of the network corresponding to parameters in the regression model, and the objective function implemented in the training of the network defining the error structure. The aim ofthe present study is to use these statistical insights to critically appraise the reliability and the precision of the predicted outputs from a trained hiddenlayer feed forward neural network. / Thesis (M.Sc.)-University of Natal, Pietermaritzburg, 2000.

In this thesis, we have considered the inference aspects of sampling from a finite population. There are significant differences between traditional statistical inference and finite population sampling inference. In the case of finite population sampling, the statistician is free to choose his own sampling design and is not confined to independent and identically distributed observations as is often the case with traditional statistical inference. We look at the correspondence between the sampling design and the sampling scheme. We also look at methods used for drawing samples. The non – existence theorems (Godambe (1955), Hanurav and Basu (1971)) are also discussed. Since the minimum variance unbiased estimator does not exist for infinite populations, a number of estimators need to be considered for estimating the same parameter. We discuss the admissible properties of estimators and the use of sufficient statistics and the Rao-Blackwell Theorem for the improvement of inefficient inadmissible estimators. Sampling strategies using auxiliary information, relating to the population, need to be used as no sampling strategy can provide an efficient estimator of the population parameter in all situations. Finally few well known sampling strategies are studied and compared under a super population model. / Thesis (M.Sc.)-University of KwaZulu-Natal, Westville, 2007.

The purpose of this study is to investigate and understand data which are grouped into categories. At the onset, the study presents a review of early research contributions and controversies surrounding categorical data analysis. The concept of sparseness in a contingency table refers to a table where many cells have small frequencies. Previous research findings showed that incorrect results were obtained in the analysis of sparse tables. Hence, attention is focussed on the effect of sparseness on modelling and analysis of categorical data in this dissertation. Cressie and Read (1984) suggested a versatile alternative, the power divergence statistic, to statistics proposed in the past. This study includes a detailed discussion of the power-divergence goodness-of-fit statistic with areas of interest covering a review on the minimum power divergence estimation method and evaluation of model fit. The effects of sparseness are also investigated for the power-divergence statistic. Comparative reviews on the accuracy, efficiency and performance of the power-divergence family of statistics under large and small sample cases are presented. Statistical applications on the power-divergence statistic have been conducted in SAS (Statistical Analysis Software). Further findings on the effect of small expected frequencies on accuracy of the X2 test are presented from the studies of Tate and Hyer (1973) and Lawal and Upton (1976). Other goodness-of-fit statistics which bear relevance to the sparse multino-mial case are discussed. They include Zelterman's (1987) D2 goodness-of-fit statistic, Simonoff's (1982, 1983) goodness-of-fit statistics as well as Koehler and Larntz's tests for log-linear models. On addressing contradictions for the sparse sample case under asymptotic conditions and an increase in sample size, discussions are provided on Simonoff's use of nonparametric techniques to find the variances as well as his adoption of the jackknife and bootstrap technique. / Thesis (M.Sc.)-University of Natal, Durban, 1998.

The objectives of this thesis is to explore different approaches of modelling clustered correlated data in the form of repeated or longitudinal counts data leading to a replicated Poisson process. The specific application is from repeated epileptic seizure time to events data. Two main classes of models will be considered in this thesis. These are the marginal and subject or cluster specific effects models. Under the marginal class of models the generalized estimating equations approach due to Liang and Zeger (1986) is first considered. These models are concerned with population averaged effects as opposed to subject-specific effects which include random subject-specific effects such that multiple or repeated outcomes within a subject or cluster are assumed to be independent conditional on the subject−specific effects. Finally we consider a distinct class of marginal models which include three common variants namely the approach due to Anderson and Gill (1982), Wei et al (1989) and Prentice et al. (1981) / Thesis (M.Sc.)-University of KwaZulu-Natal, Pietermaritzburg, 2010.