21623 research outputs found
Sort by
Transitioning from school to work: a narrative inquiry of the experiences of out-of school youth with disabilities who attended Newton Pre-Vocational School in the uMgungundlovu District Municipality.
No full textMasters Degree. University of KwaZulu-Natal, Pietermaritzburg.A narrative enquiry was conducted to explore the experiences of out-of-school youth with disabilities who attended Newton Pre-Vocational School in the uMgungundlovu District Municipality of South Africa. Semi-structured interviews and a focus group interview were used to obtain the data for this study. The study adopted a qualitative research approach, located within a critical research paradigm. Six participants who were former learners at Newton Pre-Vocational School were selected for the study; their ages ranged from 19 to 25. Findings revealed that although participants reported that their schooling experience at Newton Pre-Vocational School was much better than their primary schooling, 5 of the 6 participants felt that the Technical Occupation Curriculum used at the school did not prepare them adequately for adulthood and for their chosen field of work. In addition, as the programme was not recognised, the certificate that they earned from Newton did not open doors for them when they applied for jobs. The study found that 5 of the 6 of the participants had attempted more than once to obtain employment. Participants described companies as lacking understanding of their disability; as a result, they experienced prejudice and discrimination. This contributed negatively to participants already compromised mental wellbeing. The societal contribution to their mental state cannot be underestimated. It is imperative that stakeholders at all levels of society—parents, schools, the basic and higher education departments, the employment sector and department of Social Development—cooperate closely to support learners with special education needs to be able to transition successfully from school to the world of work and to contribute economically and socially. Further research exploring how stakeholders can support people living with disabilities—especially individuals with mild to moderate intellectual disabilities—as they transition from schooling to employment, is necessary
Exploring learners’ understanding of environmental issues: narratives of grade 7 learners in a rural uMgungundlovu district.
No full textMasters Degree. University of KwaZulu-Natal, Pietermaritzburg.Water scarcity, pollution, deforestation, and poor infrastructure represent critical environmental challenges facing developed and developing nations worldwide. These issues are particularly pronounced in rural communities, where inequalities persist between urban and rural populations. In South Africa’s uMgungundlovu Education District, such challenges are exacerbated by socio-economic disparities. Despite limited resources for disseminating information, children in rural areas fundamentally understand environmental issues and their implications for human survival. This study adopts a qualitative narrative inquiry approach within the critical paradigm to explore children’s perceptions of environmental challenges in rural settings. Data collection involved semi-structured interviews, focus group discussions, and mapping exercises with eight purposively selected participants. Findings indicate that children possess a nuanced awareness of environmental issues, recognising the impact of deforestation and pollution on the natural ecosystem and human well-being. They also demonstrate an understanding of societal inequities, including marginalisation and exclusion experienced by rural communities. Key factors contributing to environmental degradation include the unsustainable use of natural resources, such as wood for fuel, and inadequate waste management practices. Moreover, poor infrastructure, particularly during rainy seasons, hinders access to essential services like education and healthcare. Power dynamics between political entities and private corporations further complicate addressing these challenges, often resulting in incomplete infrastructure projects. The study highlights and elevates the importance of collaborative efforts among schools, communities, government agencies, and external stakeholders to promote environmental awareness and advocate for environmental justice in rural areas. It emphasises children’s agency as active community participants, deserving recognition and respect for their perspectives and experiences. Ultimately, the findings highlight the need for sustained environmental campaigns and projects to foster conservation efforts and ensure equitable access to a conducive environment
Performance comparison between gas generator and electrically pumped rocket engines under ablative and regenerative cooling.
No full textMasters Degree. University of KwaZulu-Natal, Durban.In recent years, there has been a dramatic advancement in the global satellite industry, with new technologies allowing for smaller and more powerful satellites to be developed. Despite this, South Africa and other African countries developing satellite technologies still depend on foreign launch services as there is no indigenous capability in Africa, incurring additional costs and delays. Against this backdrop, the University of KwaZulu-Natal’s (UKZN) Aerospace Systems Research Institute (ASRI) is pursuing the development of a two-stage Commercial Launch Vehicle (CLV) to provide South Africa with a sovereign launch capability. ASRI is currently developing the booster engine for CLV - the South AFrican FIrst Rocket Engine (SAFFIRE). The booster stage will utilize a cluster of nine of the SAFFIRE engines, and the second stage will use a vacuum derivative of the SAFFIRE engine called SAFFIRE-V and will give CLV the envisioned payload capacity of 200 kg to a 500 km Sun Synchronous Orbit (SSO). Although the design of the SAFFIRE engine is largely complete in terms of injector design, combustion chamber and nozzle geometry, and thrust output, the feed cycle responsible for delivering the propellants to the combustion chamber is undecided. In this report, the gas generator and electropump cycles are under consideration to supply propellants to the SAFFIRE and SAFFIRE-V chambers. In addition, although the chambers are currently ablatively cooled, ASRI remains interested in the possibility of using regenerative cooling. This study, therefore, considers the gas generator and electropump cycles under both cooling methods to identify which SAFFIRE engine combination can provide the best performance for CLV. To measure the performance of each engine combination, the payload mass capabilities of a hypothetical CLV are calculated using results attained from 1- dimensional simulations run in Flownex®, models coupled with Mass Estimating Relationships (MER) that evaluate the components that make up a two-stage liquid rocket engine. The performance comparison found that each engine configuration exceeded the set payload capacity of 200 kg. For the electropump cycle, the ablative and regenerative engines achieved payload masses of 303 kg and 290 kg, respectively. The gas generator performed even better due to a lower dead mass than the electropump cycle, achieving payload masses of 392 kg and 386 kg, respectively. For both cycles, the ablatively cooled rocket engines had better payload capabilities than the regeneratively cooled engines due to a regeneratively cooled engine having a smaller expansion ratio. When the expansion ratios were made the same, the regeneratively cooled engines achieved a payload capacity of 308 kg and 405 kg for the electropump and gas generator cycles, respectively, due to the increased thrust from the heated fuel, which has an increased energy density, producing more thrust. When the dead mass of the electropump cycle was decreased by ejecting the depleted battery packs for each engine stage during the launch, the electropump payload deficit decreased from 98 kg to 29 kg for the ablative engines and 96 kg to 32 kg for the regenerative engines. Based on this study, the best performing SAFFIRE engine configuration is a gas generator cycle with ablative cooling, giving the conceptual CLV rocket a payload of 392 kg to 500 km Sun Synchronous Orbit
Genetic analysis and hybrid prediction in tropical maize (Zea mays L.) using phenotypic and single nucleotide polymorphic markers.
No full textDoctoral Degree. University of KwaZulu-Natal, PietermaritzburgMaize (Zea mays L., 2n = 2x = 20) is a commodity crop serving the food, feed, and processing industries globally. The productivity of maize in Africa remains low (< 2 t/ha) due to various yieldlimiting factors, including abiotic stresses (such as drought, heat stress, flooding, waterlogging, erosion and poor soil health), and biotic stresses (e.g. foliar diseases and insect pests). Limited adoption of new high yielding varieties, slow rate of varietal turnover , socio-economic constraints, and policy issues further hinder productivity. Seed Co Limited is a Pan-African seed company involved in the research, development, and commercialization of seeds of major food security grain crops, including maize. The Seed Co breeding program aims to enhance the yields of new generation maize cultivars via hybrid breeding by utilizing complementary and contrasting inbred lines. New lines and experimental hybrids are developed and phenotyped using economic agronomic traits and genotyped using high-resolution Single Nucleotide Polymorphism (SNP) markers to facilitate effective selection. Integrating phenotypic and genomic selection accelerates the development of inbred lines with desirable traits to create high-performing single crosses and three-way hybrids. The new hybrids should undergo rigorous field testing for yield gains and stability across various locations to guide cultivar release and commercialization. Therefore, to complement this breeding initiative, the objectives of the study were: to assess a maize germplasm panel's genetic diversity and population structure comprising 182 founder lines and 866 derived inbred lines using Single Nucleotide Polymorphism (SNP) markers to identify genetically unique lines for hybrid breeding, to conduct genome-wide prediction of yield and component traits using qualitative and quantitative phenotypic traits and SNP markers based on the additive-dominant genomic best linear unbiased predictions model to compute genomic estimated breeding values and genomic estimated genetic values to guide inbred line development and hybrid breeding, to assess the gains in yield and yield components among single cross maize hybrids selected through genomic prediction across representative locations to guide breeding and production and to determine the combining ability effects of newly selected inbred lines and quantify the magnitude of heterosis and genotype by environmental interaction (GEI) effects of single cross hybrids to select and recommend contrasting elite lines and experimental hybrids.
In the first study, 182 founder and 866 derived maize inbred lines were characterized for genetic diversity and population structure analyses using SNP markers to identify genetically unique lines for hybrid breeding through beneficial allelic combinations. Genotyping was performed using the Affymetrix platform for the 182 founder lines (1201 SNP markers) and the Midseq platform for the 866 derived lines (1484 markers). Moderate genetic variation with genetic distance ranging from 0.004 to 0.44 (mean: 0.25) for founder lines and 0.004 to 0.34 (mean: 0.13) for derived lines was observed. Heterozygosity values ranged from 0.00 to 0.24 for both lines. About 82% of the 1201 markers and 84% of the 1484 markers exhibited polymorphism information content ranging from 0.25 to 0.50, detecting a high level of genetic diversity and that the SNPs were highly informative in distinguishing the tested lines. Analysis of molecular variance revealed significant genetic differences (P ≤ 0.001) among and within populations in the founder and derived lines. Notably, within-population variations accounted for 97% (founder lines) and 88.38% (derived lines) of the detected variations. Population structure analysis identified three subpopulations among founder lines and two among derived lines, which was supported by cluster analysis. Based on pairwise comparisons, genetically distant lines were selected, including G15NL337 and G15NL312 (Cluster 1), 15ARG152 and RGS-PL44 (Cluster 2), RGS-PL44 and 15ARG149 (Cluster 2), and RGS-PL33 and RGS-PL44 (Cluster 2). The selected lines are genetically distinct and recommended for marker-assisted hybrid maize breeding to leverage beneficial alleles.
The second study genotyped 1,102 genetically diverse inbred lines from two heterotic groups (N3 and SC) using high-density SNP markers. The 1,102 lines and 4 testers were crossed in a line-by-tester design to generate 2,830 single cross hybrids (SCHs). Phenotypic data were collected from field trials with the following SCHs: 684 evaluated at five locations in 2018/19, 760 at four locations (2019/20), 646 at four locations (2020/21), and 740 at four locations (2021/22) summer seasons in Zimbabwe. The trials were laid out in a 6 x 7 alpha lattice design with two replications at each site. 20 highperforming and contrasting inbred lines with the highest genomic estimated breeding values (GEBVs) and genomic estimated genetic values (GEGVs), each from the two heterotic groups, were identified for genetic advancement, combining ability tests and commercial hybrid development. 20 highperforming candidate SCHs with high GEGVs were identified for three-way hybrid development, variety registration and commercialization.
In the third study, 30 SCHs were developed from 11 inbred lines (6 from the N3 group and 5 from the SC group) with the highest predicted GEGVs for grain yield and associated traits using the genotypic best linear unbiased prediction (GBLUP) model. The lines were crossed using a factorial mating design with the six N3 lines used as female and five SC lines as male. The derived 30 SCHs and six commercial single cross check hybrids were field evaluated in seven locations, four in Zimbabwe and three in Zambia using a 6 x 6 alpha lattice design with two replications at each location. A combined analysis of variance revealed significant (P≤0.05) variation among the hybrids for the assessed 11 quantitative traits. Significant yield gains were realized over the mean of checks (at 13.09%), mean of the population (10.83%) and mean of best check (1.47%). Moderate to high broad-sense heritability (50 to 94%) and genetic advance were recorded for most of the assessed traits, indicating the success of selection assisted by genomic predictions. The study identified three best single cross hybrids (i.e., CTL03 x G16NL721, CTL03 x G17NL544 and GS-PL07 x G17NL544) with high and stable yields and recommended for commercialization.
In the fourth study, 11 elite inbred lines (6 female parents from N3 and 5 male parents from SC group) were crossed using a factorial mating design, resulting in 30 SCHs. The lines were selected based on the highest GEGVs for yield and component traits through GS using the GBLUP model. The 30 SCHs and six commercial check hybrids were field evaluated at seven locations (four in Zimbabwe and three in Zambia) during the 2022/2023 summer season. The trials were arranged in a 6 x 6 alpha lattice design with two replications at each location. Data were recorded on yield and yield components, and general combining ability (GCA) and specific combining ability (SCA) effects were computed. Significant GCA effects for grain yield (GY) were noted for lines CTL03, G17NL544, G16NL721, and GS-PL07, while significant SCA effects were recorded for crosses 15AG163 x G16NL679, G15NL304 x G17NL642, and 15AG162 x G16NL679. The additive main effects and multiplicative interaction (AMMI) model explained 38.95%, 50.58% and 7.24% of the total variation in GY due to genotype (G), environment (E), and genotype x environment interaction (GEI) effects in that order. The test locations were clustered into two mega environments: Rattray Arnold Research Station (RARS), Agricultural Research Trust (ART), Mpongwe Research Station (MPRS), and Lusaka West Research Station (LWRS) (Environment 1), and Mkushi Research Station (MKRS), Stapleford Research Centre (STAP), and Kadoma Research Centre (KRC) (Environment 2). The genotype and genotype-by-environment interaction (GGE) biplot analysis identified hybrids G15NL304 x G17NL544 and 15AG162 x G17NL544 as high-yielding and stable, suitable for commercialization. The two mega-environments and the selected stable, high-yielding general and specific combiners are recommended for genotype evaluation and production in Zimbabwe, Zambia, and comparable agroecologies.
Overall, the present study identified contrasting and genetically delineated inbred lines and enhanced the existing heterotic groups using high-throughput SNP markers. Best-performing lines (e.g. CTL03 and GS-PL07) were selected from the N3 heterotic group and G17NL544 and G16NL721 from the SC heterotic group. New single cross hybrids, such as CTL03 x G16NL721, CTL03 x G17NL544, and GS-PL07 x G17NL544, were selected with grain yields of 8.38 t/ha, 8.24 t/ha, and 8.23 t/ha, respectively. The new experimental hybrids are recommended for three-way hybrid development or release following multi-environment evaluation
Challenges and prospects of the national indigenous knowledge systems policy in integrating African traditional medicines into the public healthcare system in South Africa.
No full textDoctoral Degree. University of KwaZulu-Natal, Durban.Using a mixed method approach, the study investigated the challenges and prospects of the National Indigenous Knowledge Systems Policy (2004) in integrating African Traditional Medicines into the
public healthcare system in South Africa. African Traditional Medicines (ATM) and healing systems are increasingly recognized as an important aspect of the primary public healthcare delivery system
within and outside South Africa. This is primarily the case in predominantly rural and marginalized communities with limited conventional healthcare services. ATM and healing systems are integral to
African cultures and local knowledge systems. The study used a comparative case study and participatory approach, focusing on the uGu and uMkhanyakude District Municipalities in KwaZulu-Natal Province, South Africa. The study aimed to understand the knowledge and perceptions of local communities as healers and clients, regarding the prospects and challenges of the implementation of the National Indigenous Knowledge Systems (IKS) Policy (2004), with special reference to ATM. The study followed a comparative approach of two district municipalities in the KwaZulu-Natal Province, South Africa, with different ecological systems, i.e. uGu and uMkhanyakude District Municipalities. The comparative approach was chosen due to ATMs' cultural and ecological specificity, with uMkhanyakude District in the north of the province having an arid environment and uGu in the south being tropical. These ecological differences were considered critical in assessing traditional medicinal knowledge and healing practices. The study suggested that these ecological differences must be considered when implementing the IKS Policy (2004). Findings revealed that ATM use was prevalent in rural and marginalized communities of both district municipalities, mainly due to limited conventional healthcare services, and the affordability, accessibility, and cultural acceptability of ATM, especially among marginalized communities and social groups such as women, children, and the elderly. However, the majority of traditional healthcare practitioners and their clients in both district municipalities were not ware of the existence of the National IKS Policy (2004). The study recommended the following: 1. Because substantial numbers of ordinary people in African local communities, including the study areas, onsulted ATHPs for primary healthcare, this study suggests the great need for more comparative, culturally and ecologically specific research studies to understand the significance of this healthcare and associated local community-based knowledge systems in advancing healthcare, social and epistemic justice. 2. The limited knowledge and awareness among various stakeholders, including ATHPs, regarding policy frameworks related to IKS and ATM, calls for a deeper investigation of specific challenges commonly encountered by ATHPs and their clients across South Africa. This investigation should also explore the implications for policy development and implementation. 3. Finally, a critical review of existing legislation and active engagement with relevant policymakers to address the challenges of integrating African traditional medicine into the public healthcare system is recommended for future studies
Discontinuity without change? the place and discourse of colonial memory in Zimbabwe’s post- Mugabe Zanu-PF politics.
No full textDoctoral Degree. University of KwaZulu-Natal, Durban.Zimbabwean politics are notably complex and difficult to understand, even by scholars with a strong interest in African affairs with a long institutional memory of the historical determinants of the independence and post-independence struggles within Zimbabwe. Through the lens of political culture and functional theory campaign communication, this qualitative inquiry titled “Discontinuity without change? The place and discourse of colonial memory in Zimbabwe’s post-Mugabe ZANU-PF politics” scrutinizes the colonial narratives in the political discourses in Zimbabwe’s ruling party ZANU (PF) following the Robert Mugabe era intending to understand how colonial memory shapes the party’s the ideological foundations and policy directions. The study draws on eight speeches delivered by former president Robert Mugabe during the 2002 elections, as well as speeches by his successor and current president Emmerson Munangagwa during the 2018 election campaign. It also incorporates insights from key informants within ZANU (PF), Zimbabwe Broadcasting Corporation (ZBC), Zimpapers, and Alpha Media Holdings (AMH) to explore the role of media in influencing the nuanced interplay between historical legacies, political discourse, and contemporary governance. By examining the ebbs and tides of electoral politics in Zimbabwe spanning nearly decades through the prism of post-colonial memory, the study concludes that while Mnangagwa’s ascendance as President hinted at a departure from his predecessor’s politics, there is a notable continuity in the streams of colonial memory that informed ZANU-PF electoral strategies. This underscores how political discourses and power dynamics during elections are deeply entrenched within the broader context of Zimbabwean politics and pan-African pursuit of of self-determination (Nyika inovakwa nevene vayo), identity and independence. Despite certain shifts in Mnangagwa’s ‘New Dispensation’ that deviate from Mugabeism, the persistence of colonial memory underscores its pivotal role in shaping the principles and practices of representative democracy within Zimbabwe. The media’s influence in (re)shaping post-Mugabe discourse sheds light on the implications of memory appropriation in contemporary Zimbabwean political communication
Empathy, remorse, and restoration of dignity contributing to reduced recidivism: assessing the role of restorative justice in promoting offender rehabilitation and reintegration in Durban.
No full textResearch Articles. Criminology and Forensic Studies.This study aims to determine whether restorative justice (RJ) programmes can help
reduce recidivism and facilitate the effective reintegration of offenders into society by
encouraging empathy, regret, and the restoration of their dignity. The research
investigates the lived experiences of key informants and community members in Cato
Manor, South Africa, using a qualitative methodology approach. The significance of
restorative justice (RJ) programmes in fostering possibilities for regret among offenders
and a feeling of community connection was demonstrated using thematic analysis.
These results support the Social Bond Theory (SBT) and Reintegrative Shaming Theory
(RST), indicating that restorative justice (RJ) can promote accountability and lower
recidivism rates. Families must take an active role in helping offenders get back on
track and reintegrate into society through therapy and support services. To guarantee
the efficacy of RJ programmes and to ensure that they address the unique requirements
of the community, the study highlights that there is a necessity of involving community
members in their design and evaluation. RJ programmes can provide a route to a more
efficient and compassionate judicial system by emphasising communication, healing,
and reintegration; this will ultimately lead to safer and more equitable societies. The
study’s conclusions advocate for a comprehensive strategy to lower recidivism, one that
includes active community involvement in RJ efforts, family assistance, focused
interventions, and career counselling for ex-offenders
Forest image classification based on deep learning and ontologies = Ukwahlukaniswa kwesithombe sehlathi ngokusekelwe ekufundeni okujulile kanye nobunjalo bolwazi.
No full textDoctoral Degree. University of KwaZulu-Natal, Pietermaritzburg.Forests contribute abundantly to nature’s natural resources and they significantly contribute to a wide range of environmental, socio-cultural, and economic benefits. Classifications of forest vegetation offer a practical method for categorising information about patterns of forest vegetation. This information is required to successfully plan for land use, map landscapes, and preserve natural habitats. Remote sensing technology has provided high spatio-temporal resolution images with many spectral bands that make conducting research in forestry easy. In that regard, artificial intelligence technologies assess forest damage. The field of remote sensing research is constantly adapting to leverage newly developed computational algorithms and increased computing power. Both the theory and the practice of remote sensing have significantly changed as a result of recent technological advancements, such as the creation of new sensors and improvements in data accessibility. Data-driven methods, including supervised classifiers (such as Random Forests) and deep learning classifiers, are gaining much importance in processing big earth observation data due to their accuracy in creating observable images. Though deep learning models produce satisfactory results, researchers find it difficult to understand how they make predictions because they are regarded as black-box in nature, owing to their complicated network structures. However, when inductive inference from data learning is taken into consideration, data-driven methods are less efficient in working with symbolic information. In data-driven techniques, the specialized knowledge that environmental scientists use to evaluate images obtained through remote sensing is typically disregarded. This limitation presents a significant obstacle for end users of Earth Observation applications who are accustomed to working with symbolic information, such as ecologists, agronomists, and other related professionals. This study advocates for the incorporation of ontologies in forest image classification owing to their ability in representing domain expert knowledge. The future of remote sensing science should be supported by knowledge representation techniques such as ontologies. The study presents a methodological framework that integrates deep learning techniques and ontologies with the aim of enhancing domain expert confidence as well as increasing the accuracy of forest image classification. In addressing this challenge, this study followed the following systematic steps (i) A critical review of existing methods for forest image classification (ii) A critical analysis of appropriate methods for forest image classification (iii) Development of the state-of-the-art model for forest image segmentation (iv) Design of a hybrid model of deep learning and machine learning model for forest image classification (v) A state-of-the-art ontological framework for forest image classification. The ontological framework was flexible to capture the expression of the domain expert knowledge. The ontological state-of-the-art model performed well as it achieved a classification accuracy of 96%, with a Root Mean Square Error of 0.532. The model can also be used in the fruit industry and supermarkets to classify fruits into their respective categories. It can also be potentially used to classify trees with respect to their species. As a way of enhancing confidence in deep learning models by domain experts, the study recommended the adoption of explainable artificial intelligence (XAI) methods because they unpack the process by which deep learning models reach their decision. The study also recommended the adoption of high-resolution networks (HRNets) as an alternative to traditional deep learning models, because they can convert low-resolution representation to high-resolution and have efficient block structures developed according to new standards and they are excellent at being used for feature extraction.
Iqoqa.
Inkambu yocwaningo lwezinzwa ezikude ihlala ivumelana nezimo ukuze kuthuthukiswe ama-algorithms ekhompuyutha asanda kuthuthukiswa kanye namandla ekhompuyutha akhulayo. Kokubili ithiyori kanye nokwenza kokuzwa okukude kushintshe kakhulu ngenxa yentuthuko yakamuva yezobuchwepheshe, njengokwakhiwa kwezinzwa ezintsha kanye nokuthuthukiswa kokufinyeleleka kwedatha. Izindlela eziqhutshwa idatha, okuhlanganisa abahlukanisa izigaba abagadiwe (njengaMahlathi Angahleliwe) kanye nezigaba zokufunda ezijulile, zibonise ukunemba ekudaleni izithombe ezibonakalayo futhi ngaleyo ndlela zithola ukubaluleka okukhulu ekucubunguleni idatha enkulu yokubhekwa komhlaba. Nakuba amamodeli okufunda ajulile ekhiqiza imiphumela egculisayo, abacwaningi bakuthola kunzima ukuqonda ukuthi benza kanjani izibikezelo ngenxa yemvelo yabo yebhokisi elimnyama eliphuma ezinhlakeni zenethiwekhi eziyinkimbinkimbi ngokwemvelo. Lolu cwaningo lukhuthaza ukufakwa kobunjalo bolwazi (ontologies) ekuhlukaniseni izithombe zehlathi ngenxa yekhono lakho lokumela ulwazi lochwepheshe besizinda. Ucwaningo luveza uhlaka lwe-methodological oluhlanganisa amasu okufunda ajulile kanye nobunjalo bolwazi ngenhloso yokusebenzisa ubuchwepheshe besizinda nokukhulisa ukunemba kokuhlukaniswa kwezithombe zehlathi. Uhlaka lobunjalo bolwazi lwaluguquguquka ukuze luthwebule ukubonakaliswa kolwazi lochwepheshe besizinda. Imodeli yesimanjemanje yobunjalo bolwazi yenze kahle njengoba ithole ukunemba kwezigaba okungama-96%, nge-Root Mean Square Error engu-0.532. Imodeli inganwetshwa nasembonini yezithelo nezitolo ezinkulu ukuze zihlukanise izithelo ngezigaba zazo. Ingase futhi isetshenziselwe ukuhlukanisa izihlahla ngokuphathelene nezinhlobo zazo. Njengendlela yokuthuthukisa ukuzethemba kumamodeli okufunda okujulile ngochwepheshe besizinda, ucwaningo luncome ukwamukelwa kwezindlela zobuhlakani bokwenziwa ezichazwayo (i-XAI) ngoba ziveza inqubo lapho amamodeli okufunda okujulile afinyelela khona ezinqumweni zawo. Ucwaningo luphinde lwancoma ukwamukelwa kwamanethiwekhi anesinqumo esiphezulu (HRNets) njengenye indlela yamamodeli okufunda okujulile ngendabuko ngoba angaphambili angaguqula ukumelwa kokucaca okuphansi kube ukucaca okuphezulu. I-HRNets futhi inezakhiwo zamabhulokhi ezisebenza kahle ezakhiwe ngokuvumelana nezindinganiso ezintsha futhi zibonise ukukhishwa kwezici ezisebenzayo. Ucwaningo luncoma ukuthi ikusasa lesayensi yezinzwa ezikude kufanele lisekelwe amamodeli okufunda okujulile asezingeni eliphezulu ahambisana namasu okumela ulwazi lwesizinda njengokobunjalo bolwazi
Stable distributions with applications to South African financial data.
No full textDoctoral Degree. University of KwaZulu-Natal, Durban.In recent times, researchers, analysts and statisticians have shown a keen interest in studying Extreme Value Theory (EVT), particularly with the application to mixture models in the medical and financial sectors. This study aims to validate the use of stable distributions in modelling three Johannesburg Stock Exchange (JSE) market indices, namely the All Share Index (ALSI), Banks Index and the Mining Index, as well as the United States of American Dollar (USD) to South African Rand (ZAR) exchange rate. This study leverages the unique properties of stable distributions when modelling heavy-tailed data. Nolan’s S0-parameterization stable distribution (SD) was fitted to the returns of the three FTSE/JSE indices and USD/ZAR exchange rate and a hybrid Generalized Autoregressive Conditional Heteroskedasticity (GARCH)-type model combined with stable distributions was fitted to each return series. The two-tailed mixture model of the Generalized Pareto
Distribution (GPD), stable distribution, Generalized Pareto Distribution referred to as GSG, as well as the Stable-Normal-Stable (SNS) and Stable-KDE-Stable (SKS) was fitted to evaluate its relative performance in modelling financial data. Results show that the S0-parameterization SD fits the South African financial returns well. The hybrid GARCH (1,1)-SD model competes favourably with the GARCH-GPD model in estimating Value-at-Risk (VaR) for FTSE/JSE Banks Index, FTSE/JSE Mining Index and the USD/ZAR exchange rate returns. The hybrid EGARCH (1,1)-SD competes
well against the GARCH-GPD model for the FTSE/JSE ALSI returns. Inconclusive results are observed for the short position of the fitted GKG models; however, in the long position, an appropriate fit of the GPD-KDE-GPD (GKG) model, where KDE is the kernel density estimator, is emphasised for all four return series. The proposed mixture models, GSG, SNS and SKS models, are found to be a good alternative in fitting South African financial data to the commonly used GPD-Normal-GPD (GNG) mixture model. The results of this study are important to financial practitioners, risk managers and researchers as the proposed mixture models add more value to the
literature on the applications of extreme mixture models.Author's Keywords: Stable distributions, Nolan’s S0-parameterization, mixture models,
GPD-Normal-GPD, GPD-Stable-GPD, Stable-Normal-Stable, Stable-KDE-Stable,
Kolmogorov-Smirnov test, Anderson-Darling test, VaR, Kupiec likelihood rati
Investigation of the relationships between host genetics and COVID-19 disease progression among different ethnic groups in South Africa.
No full textDoctoral Degree. University of KwaZulu-Natal, Durban.Abstract available in PDF.Abstract part of Chapter 1