Masters

Permanent URI for this collection

https://open.uct.ac.za/handle/11427/29120

Browse

Recent Submissions

Now showing 1 - 5 of 21402
  • No Thumbnail Available
    Item
    Open Access
    The non-parametric calibration of jump-diffusion lévy models
    (2022) Van Zyl, jaco; Ouwehand, Peter
    This study investigates the effectiveness of relative-entropy-based regularised calibration procedures at addressing the ill-posedness encountered in the calibration of nonparametric jump-diffusion Lévy models. Calibrated models, which have been selected for their capability of simulating realistic price path evolutions, are typically employed to price path-dependent instruments or to perform dynamic hedging. The financial risks associated with using pricing models in real-world transactions can be reduced by selecting an appropriate model together with a suitable calibration procedure. It has been well established that mispricing resulting from improper modelbased pricing has resulted in significant financial losses. Lévy processes aim at improving on the ability of continuous models to represent market-observed price evolutions by allowing highly flexible jump characteristics to be specified, enabling the inclusion of jump-discontinuities in pricing models. The capacity of parametric Lévy models to provide a general class of feature-rich models with jump-discontinuities is further enhanced by the extension to non-parametric Lévy models, where the jump characteristics can more freely be defined. Model calibrations, like most other inverse problems, suffer from ill-posedness. Consequently, optimisation results from a calibration exercise generally do not converge to a unique solution and typically do not vary continuously with input data. The increase in dimension of the solution space associated with the non-parametric approach necessitates further measures to address the ill-posedness. This study evaluates the performance of the relative-entropy-based regularised calibration procedures proposed by Cont and Tankov [CT04, CT06] at addressing the primary concern of ill-posedness encountered in the calibration of non-parametric jump-diffusion Lévy models. We will show that although the procedures provide some stability with respect to the input prices between subsequent calibrations, the procedures are of limited value at addressing the ill-posedness relating to the convergence to a unique solution. In our experiments, we expose the sensitivity of results to both the initial points as well as to the prior measure presented to the optimisation procedures. Our study highlights some deficiencies in the process, showing that the regularised calibration procedure is unreliable, necessitating active user intervention to manage outcomes. We conclude that the observed difficulties are primarily the result of a persistent non-convexity of the regularised objective function at realistic levels of regularisation. Therefore ill-posedness continues to present a risk that needs to be managed by practitioners when applying these procedures to the recovery of non-parametric Lévy models.
  • No Thumbnail Available
    Item
    Open Access
    Using working memory to predict other domains within the learner profiler in an older adolescent sample
    (2023) Petersen, Asheeqa; Schrieff-Brown, Leigh
    Learning difficulties and disabilities (LDD) are the most frequently diagnosed of childhood developmental disorders. In South Africa (SA), however, a standard and nationally accepted tool has not yet been established for assessing LDDs and thus, specific incidence rates are not known. An underlying factor which may be important to consider in the context of LDDs is working memory (WM) which has been reported as playing a crucial role in learning and WM deficits appear to be higher in the context of LDDs. Thus, it may be imperative to adopt, and adapt to, new technologies that are both cost-effective and easily accessible, addressing the gap in resource availability. The Learner Profiler (LP) is an example of one such computerised test, being relatively cost effective and accessible. It should be noted, however, that research on the LP test method is particularly limited due to both a scarcity in the literature and the use of small sample sizes in said literature. The aim of this research was to investigate whether a computerised tool of WM on the LP could predict scores on other computerised cognitive domains on the LP. This quantitative study used a within-subjects experimental design to investigate the predictive value of LP WM module in other LP modules, namely, Visuospatial, Spelling, Missing Word, Word Choice, and New Word Spelling. The sample comprised of 1175 participants aged between 16 to 19 years old. At the time, participants attended a Technical Vocational Education and Training college (TVET) situated in an urban area in Gauteng, SA. The LP modules were administered during normal admissions processes at the college. To assess the predictive value of the LP WM module, the scores of the modules were analysed using multiple regression analyses.
  • No Thumbnail Available
    Item
    Open Access
    Agile project management in South African financial service organisation: a case study
    (2023) Mhlanga, Success Siphesihle; Rivett, Ulrike
    Financial service organisations have traditionally utilised conventional project management approaches to execute software projects. However, with the emergence of the agile methodology, there has been a growing transition among these organisations towards adopting agile project management (APM) practices. Scholars in the field have pointed out that traditional project management approaches are inadequate in meeting the dynamic demands of the financial service sector. This observation helps to explain the industry's inclination towards alternative approaches. The increasing trend of organisations embracing Agile Project Management (APM) highlights a pressing need to rethink the delivery mechanisms for software development projects. Previous studies have focused on documenting employees' experiences during an agile transition, but there is a need for further examination of the experiences of management. This study analysed the perceptions of managers in a financial service organisation during an APM transition. Utilising a case study methodology, perceptions, and experiences of 14 managers were analysed using a qualitative research paradigm. The study showed that the financial service organisation transitioned to leverage the benefits of agile such as incremental delivery, reaching the market faster, gaining visibility on the product output, and increasing transparency. The findings revealed that value was immediately created by increasing visibility and transparency, meeting customer demands, and quantifying return on investment. Some managers associated “walking the agile journey together” with the different levels of management as a positive attribute towards transitioning. The study found that executive management influenced decisions and drove change throughout the transition process. Additionally, a relationship between resistance to change and the absence of a change management plan was identified. The absence of a clear and communicated change management plan contributed to frustrations in persuasion, which resulted in some employees leaving the organisation. This study suggests that further research is needed to examine the consequences of transitioning without a change management plan.
  • No Thumbnail Available
    Item
    Open Access
    Interaction between host genetics and drug therapy in influencing gut bacterial diversity
    (2023) Magadze, Mulalo; Dandara, Collet
    Background: Response to drug treatment shows inter-individual variability attributed mainly to environmental and genetic factors. It is now accepted that human genome variation and microbiome profiles play significant roles in patients' response to medication. This study aimed to investigate host genetic variation that affects an individual's microbiome profile and, ultimately, its effects on the therapeutic drug response of efavirenz (EFV) and warfarin. Efavirenz and warfarin are widely prescribed drugs for treatment of HIV and clotting disorders, respectively. Genetic polymorphisms in fucosyltransferase 2 (FUT2), Neural Cell Adhesion Molecule 1 (NCAM1), Shroom Family Member 3 (SHROOM3), Vitamin D receptor (VDR) and Lactase (LCT) were characterised and evaluated on their role on microbiome profiles and how this affected drug responses. Methods: Six hundred and forty-seven (n=647) African adults were recruited from Zimbabwe and blood for DNA extraction, stool samples for microbiome analysis and plasma for drug analyses were obtained. Genetic characterisation of SNPs, FUT2 (rs601338), NCAM1(rs17115310), SHROOM3 (rs11724031), VDR (rs1544410), VDR (rs7975232), VDR (rs731236), LCT (rs2164210) and LCT (rs7579771) was performed. Correlations were then made among genotypes and microbiome profiles on participants with different drug exposures (efavirenz and warfarin). Results: The five genes exhibited genetic variation. Baseline allele frequencies for each of the SNPs was as follows; FUT2 rs601338A (0.433), NCAM1 rs17115310G (0.233), SHROOM3 rs11724031A (0.192), VDR rs1544410T (0.255), VDR rs7975232C (0.270), VDR rs731236G (0.242), LCT rs2164210C (0.188) and LCT rs7579771T (0.292). The frequencies of the variant alleles showed differences when compared to other populations. FUT2 rs601338 was shown to have an influence on warfarin Cmax. Genera Actinomycetospora and Brevibacterium; and species Corynebacterium kroppenstedtii, Macrococcus caseolyticus and Kocuria kristinae were significantly abundant in the gut bacterial composition of FUT2 rs601338 AA genotype than the GG+GA genotypes. Bacterial order Bifidobacteriales were relatively abundant in the warfarin group iii than in the untreated control group whereas Symbiobacteriaceae was significantly plentiful in the efavirenz group than the untreated control group. Discussion: Genes that affect body colonising microbiomes, although not directly important in pharmacogenomics, influence treatment outcomes. There were some differences and similarities in the distribution of minor allele for the 8 variants studied when compared to other global populations, which could point to possible differences in microbiome profiles among individuals. Different microbiome profiles identified associated with the polymorphisms are likely to affect drug treatment outcomes, through host genetics, microbiome and drug interactions. Conclusion: These results show that African populations carry different polymorphisms which are likely to affect microbiome profiles and drug responses. Thus, when considering pharmacogenomics, it is important to take into account, microbiome profiles as these affect and are affected by therapeutic drugs.
  • No Thumbnail Available
    Item
    Open Access
    In-Silico design and verification of an extracorporeal normothermic cardiac perfusion system for use during heart transplantation procedures
    (2025) Van Den Berg, Ronald; Sivarasu, Sudesh
    Heart transplantation relies on effective donor organ preservation to ensure successful graft viability. Since the first human heart transplant in 1967 by Christiaan Barnard, organ preservation techniques have evolved from static cold storage with cardioplegic arrest to continuous blood perfusion, which enhances metabolic support and extends viable ischemic time. This study explores the development of a system for continuous myocardial perfusion to improve donor heart preservation during transplantation. Using in-silico modelling and simulation, the study defines functional requirements for a proof-of-concept system capable of achieving physiologically relevant pressure and flow waveforms necessary for sustained coronary perfusion. A cardiovascular hemodynamic simulation environment was established by adapting lumped parameter models and integrating computed tomography angiograms, facilitating both in-silico and coupled in-vitro validation analyses. This enabled the development of a bench testing model that replicated physiologically relevant coronary perfusion dynamics. The bench testing model provided critical insights for refining in-silico simulations and optimising design parameters for improved myocardial perfusion. Validation was performed through vessel-specific flow rate comparisons with computational fluid dynamics simulations. Experimental results identified a time delay in relation to the identified set of functional and control parameters when achieving target physiological pressures, informing future system optimisation. Further findings allowed for the identification of relative flow proportion exiting through the Left Circumflex and Right Major Coronary arteries and was shown to behave as a second order time derivative with respect to the inflow waveform applied to a fabricated flow phantom during testing. Similarly, the proportion of flow exiting through the Left Anterior Descending and Ramus Intermedius arteries exhibited first order time derivative behaviour in relation to the inflow signal. The resultant outcomes of testing and analysis allowed for the tuning of an embedded pump control system yielding the optimised parameter control values for proportional, integral and derivative gain of 147.74, 2.57 and -4974.96 respectively. The findings of this study establish a framework for the development of an automated continuous myocardial perfusion system, contributing to enhanced donor heart preservation strategies for clinical transplantation. .