ResearchSpace

ResearchSpace is the institutional repository of the University of KwaZulu-Natal,
unlocking knowledge, empowering impact, and preserving UKZN's research legacy.


ReseachSpace Libguide

Communities in ResearchSpace

Select a community to browse its collections.

Now showing 1 - 5 of 11

Recent Submissions

Machine learning vs traditional portfolio optimisation under different market conditions in exchange traded funds portfolios.
(2025) Mutemeri, Linah.; McCullough, Kerry-Ann Frances.; Zhou, Helper.
The rapid advancement of machine learning (ML), a branch of artificial intelligence that enables systems to learn from data, has significantly reshaped financial decisionmaking. This study examines the comparative performance of traditional and MLbased portfolio optimization methods in Exchange-Traded Funds (ETFs), which provide diversified exposure to equities, bonds, commodities, and alternative assets. ETFs’ liquidity, transparency, and cost-effectiveness make them an ideal context for testing optimization strategies. The research employs a three-stage methodology. First, a Systematic Literature Review (SLR) assesses the global application of ML techniques in portfolio optimization, highlighting models such as Random Forest (RF), Long Short-Term Memory (LSTM), and Support Vector Machines (SVM). Second, an empirical analysis uses daily return data from 33 JSE-listed ETFs spanning 2013–2023 to compare three traditional models Markowitz Model (MM), Conditional Value-at-Risk (CVaR), and Mean Absolute Deviation (MAD) with three ML models: Enhanced Radial Basis Function Artificial Neural Network (K4-RANN), Genetic Algorithms (GA), and Extreme Gradient Boosting (XGBoost). Third, event study methodology evaluates portfolio resilience during South African market shocks, namely the COVID-19 pandemic (2020–2022), the July 2021 social unrest, and the April 2022 KwaZulu-Natal floods. Performance is measured through mean return, standard deviation, and Sharpe ratio. Findings show that ML-based models generally outperform traditional approaches by generating higher risk-adjusted returns and improving diversification. Under normal market conditions, GA and K4-RANN produced portfolios with lower risk and stronger returns. During crises, XGBoost proved most effective, striking a balance between returns and volatility and demonstrating adaptability in turbulent markets. These results confirm the potential of ML models to enhance portfolio stability and resilience in both normal and distressed environments. The study concludes that portfolio managers and institutional investors should integrate ML-based optimization, particularly ensemble models like XGBoost, into dynamic allocation strategies. For policymakers and financial institutions in emerging markets, investment in ML infrastructure and data capabilities is recommended to drive innovation in asset management. Future research should investigate hybrid frameworks that merge traditional financial theory with ML to advance more comprehensive portfolio optimization approaches.
The potential of human serum S100 Calcium-binding Protein B, Glial Fibrillary Acidic Protein, Neuron-Specific Enolase, and Serum Amyloid A as biomarkers for Traumatic Brain injury in moderate and severe cohort.
(2026) Mafuika, Seke Nzau.; Lazarus, Lelika.; Harrichandparsad, Rohen.
Background: Traumatic brain injury (TBI) occurs due to a severe head injury caused by an external force during a vehicle accident, fall, domestic violence, or explosion. In sub-Saharan Africa, the prevalence of TBI is 150–170 per 100,000 individuals. It is estimated that by 2050, the prevalence of TBI in Africa will increase to 14.25±0.75 million. TBI is clinically evaluated and classified using the Glasgow Coma Score (GCS), with CT and MRI. However, GCS has many limitations. Therefore, the aim of this study was to investigate the role of serum GFAP, S100B, NSE and SAA as potential biomarkers for the detection and diagnosis of TBI. Methodology: The first phase of the experiment was the scan analysis (n=50) as recorded in the hybrid electronic medical registry, MediTech, at a large urban tertiary referral hospital in eThekwini, South Africa. The second phase utilized a multiplex immunoassay (ProcartaPlex, Thermo Fisher Scientific, 2024) to detect human serum GFAP, S100B, NSE and SAA in n=50 patients with moderate to severe TBI and n=25 control patients. Ethics approval was obtained from the Biomedical Research Ethics Committee (Ref. BREC/00002147/2020) of the University of KwaZulu-Natal. Statistical analysis was achieved using Statistical Package for the Social Sciences version 30.0.0.0 (SPSS; IBM Corporation 2025). Results: Serum GFAP and S100B concentrations were significantly elevated in patients with moderate and severe TBI compared with the control group (P < 0.001). No significant differences between the moderate and severe groups were detected for GFAP or S100B (P=0.2623, P>0.999). The serum concentrations of NSE and SAA proteins were also significantly upregulated in the moderate and severe TBI compared to the control group (p<0.0001) and (p= 0.0002), respectively. Conclusion: These findings suggest that elevated serum concentrations of GFAP, S100B, NSE and SAA are indicative of increasing severity of TBI. These findings suggest that combining GFAP, S100B, NSE, and SAA biomarkers with GCS and CT findings enhances the accuracy of TBI classification.
A comprehensive audit of household waste management practices within the Bluff area (eThekwini Municipality, South Africa) to assess sustainability considerations and to inform future intervention strategies.
(2023) Gumede, Amanda Naledi.; Bob, Urmilla.; Munien, Suveshnee.
The process of waste management has predominantly included unsustainable methods such as burying, dumping and burning. Therefore, waste generators such as residential areas need to shift waste management practices from generating and disposing to more sustainable practices that include embracing circular economy principles. The research aimed to undertake a comprehensive examination of household waste management practices in the Bluff area (eThekwini Municipality, South Africa) to assess sustainability considerations and inform future intervention strategies. The study was guided by a multi-conceptual theoretical framework, drawing from the Theory of Planned Behavior (TPB), Integrated Solid Waste Management (ISWM) approach and the circular economy perspective. The study adopted a mixed methodological approach, using quantitative surveys and key informant interviews as the primary data collection tools. A total of 400 household surveys were conducted. Six key informants were purposely selected from targeted relevant groups and organizations. The nonparametric Pearson’s chi-square test of independence was used to analyze the quantitative data which was predominantly categorical. Specifically, chi-square tests were used to examine differences across variable categories. The study found that the Bluff area is largely made up of nuclear families. The households were made up of more males than females, with a fairly youthful population. The highest levels of education attained and employment status of household members portrayed comparable distributions between the sexes, with the majority of the members of households having completed secondary school and being employed. The results further revealed a dominant reliance on piped/ tap water, flush toilets and electricity. Additionally, the results indicate that there is a concerning lack of understanding of what different waste streams are at the household level. In this regard, it was found that households did not recognize the role they play in generating different waste streams. The main household disposal method chosen for almost all generated waste streams, as well as PPE/ COVID-19-related waste streams, was dustbins, which is in line with waste collection and disposal mainly being the responsibility of the Municipality. The key findings indicate that most of the households reported not separating the various waste streams. Waste separation was significantly influenced by the highest level of education attained by respondents, monthly household income and total household size. Moreover, the majority of the respondents were not aware of what sustainable waste management is. Television and school were found to be the highest sources of information about sustainable waste management practices. Furthermore, repurposing was the main sustainable waste management practice that households currently used. Not a significant percentage of respondents had previously used and discontinued using sustainable waste management practices. The results further show that households were vastly interested in using sustainable practices such as renewable energy in future. This future interest in accessing sustainable waste management facilities/ equipment was found to be associated with sex, the highest level of education attained, age, employment status and household decision-maker. Moreover, the majority of the respondents were willing to use reducing, recycling, composting and repurposing in the future. This future willingness was significantly correlated with age, total household size, household decision-maker and monthly household income. In terms of barriers, the results reveal that households mainly did not separate waste streams due to a lack of access to the necessary facilities, the practice being time-consuming, and not seeing the need to separate various waste streams. Similarly, the discontinuation of the use of sustainable waste management practices was mainly due to these practices not being easily available and these being time-consuming to engage in. The main problems associated with the use of sustainable waste management practices were also that there was an inadequate supply of equipment, that these were not easily accessible, they were time-consuming to engage in, inconvenient to use and too expensive. It is thus evident that the government and the private sector need to make sustainable waste management equipment/ facilities and practices easily accessible to communities. The study has considered comprehensive waste management practices within a historically vulnerable and heterogeneous community. Specifically, instead of generalizing and focusing on specific waste streams, the study looked at the main residential waste streams in their entirety while considering sustainability considerations among households of different socioeconomic groups. The study further unpacked the willingness-to-adopt and pay for sustainable waste management practices and the challenges experienced that constrain the transition to more sustainable waste management practices. Moreover, the study provided a further adjustment to the TPB to consider waste management practices variables at the household level. It is recommended that the government and eThekwini Municipality strengthen initiatives aimed at improving the socio-economic standing of households and that efforts be directed at investigating alternative means of domestic waste management through collaborations with surrounding areas. Individuals and households must be made aware of the pivotal role they play in waste management and are motivated and encouraged to integrate waste management into their daily routines. Additionally, increased efforts need to be directed toward understanding and catering to the management of waste streams such as e-waste while also promoting the use of energy-efficient appliances. Waste management research and initiatives need to include children and teenagers to instill environmental awareness as early as possible. Comprehensive waste management also requires collaboration between various stakeholders including waste management officials, community members, researchers and NGOs.
Three decades of shoreline change investigation over Lake Sibaya, South Africa, using Landsat imagery and geospatial methods.
(2024) Kemp, Fallon Kagney.; Xulu, Sifiso.
Lake Sibaya has experienced a significant decline in shoreline changes over the past three decades, in part, due to land use/cover (LU/LC) changes and drier climate conditions. In response, this study examined shoreline changes of Lake Sibaya between 1986 and 2020, LU/LC changes surrounding it, and climate variability using Landsat data series, the Digital Shoreline Analysis System (DSAS) geospatial modelling approaches, and climate data. DSAS provides the unique ability to compute the rate of change statistics using a suite of metrics, but in this thesis, only two variables were used (Net Shoreline Movement (NSM) and End Point Rate (EPR)) to understand shoreline dynamics and predict the future shoreline position over the next decade. The Landsat analysis showed a reduction in Lake Sibaya’s surface area from 70.7 km2 in 1986 to 49.5 km2 in 2020, while the DSAS analysis showed that the changes occurred at a NSM of −1338 m and an EPR average of ±6 m/year. Clear, conspicuous retracted changes were observed near the planted forests of Mbazwana between the southeastern/western basins and Manzengwenya in the northeastern region of the northern arm. Shoreline projections for the future showed the lake would continue to shrink from 49.5 km2 in 2020 to 39.1 km2 by 2030. Field observations and photographs validated DSAS results of declining shoreline changes in the lake and further in the vulnerable edges. Regarding the LU/LC, the results showed an expansion of forest plantations from 155.2 km2 in 1986 to 248.3 km2 in 2020, while water bodies decreased considerably, as indicated above, with a clear increase of barren land from 129.3 to 210.5 km2 over the 34 years, mostly the accumulation of sandbars in areas previously occupied by the lakes’ water. The study established a link between the lake’s surface changes and drier climate conditions during the extreme El Niño phases. During El Niño/La Niña events, the lake surface decreased/increased, with precipitation showing a significant decrease in the 2014/2016 intense drought period. During this period, higher temperatures were in the order of 32°C, and these conditions are believed to have caused further reduction in lake water levels. Overall, this research demonstrates how LU/LC change and drier climate conditions have interacted to shape Lake Sibaya's shoreline over the past 34 years and highlight the importance of integrated management approaches to sustaining this vital ecosystem. Further research is recommended to examine the development of diverse vegetation communities in dry areas of the lake as well as the intensification of illicit forestry practices in this catchment area and how these might impact the ecological function of this water resource.
Assessing the effectiveness of sector education and training authority learnerships in promoting entrepreneurial growth: a case study of AgriSETA, Eastern Cape province, South Africa.
(2025) Nxele, Thando Raymond.; Ngwenya, Charles Tony.
This study examined the effectiveness of AgriSETA-accredited learnerships in developing technical and entrepreneurial capabilities among trainees in the Eastern Cape Province. It further investigated the extent to which these competencies were translated into sustainable post-training employment or enterprise activity. Guided by Institutional Theory, the Resource-Based View, the Functional Theory of Value, and the Unified Theory of Acceptance and Use of Technology, the study adopted an explanatory sequential mixed-methods design. A survey of 230 former learners identified patterns in competence development and early post-training outcomes, followed by semi-structured interviews with facilitators, host employers, and AgriSETA officials that explored how competencies were interpreted and applied within real-world agricultural production and training environments. The findings indicate that the learnerships consistently strengthened technical agricultural skills, which emerged as the most stable and reliably achieved outcome. Entrepreneurial and digital competencies were less consistently developed and more difficult for trainees to apply independently. Structural constraints, including limited access to finance, insecure production space, weak market linkages, and the absence of structured post-training support, significantly restricted the transition from competence to sustainable economic participation. These conditions help explain the limited predictive strength of competence variables in the regression analysis. The study contributes theoretically by integrating Institutional Theory, the Resource-Based View, the Functional Theory of Value, and UTAUT to explain how capability, opportunity and support interact within vocational training systems. Methodologically, it demonstrates the value of mixedmethods designs for identifying how structural conditions shape learner trajectories in rural and resource-constrained settings. Empirically, it provides rare, context-specific evidence on the outcomes of agricultural learnerships in the Eastern Cape Province. Practically, it offers recommendations for strengthening curriculum integration, facilitator preparation, institutional coordination, and post-training support mechanisms. A framework is proposed to illustrate how technical competence must align with opportunity structures and ongoing support for skills development programs to yield meaningful and sustained post-training progression.