KovsieScholar Repository

KovsieScholar is the University of the Free State’s open access institutional repository. It preserves, showcases, and provides access to the University's scholarly and research outputs, including theses, dissertations, publications, and other digital collections, ensuring their long-term visibility, accessibility, and impact.

For assistance, contact: openaccess@ufs.ac.za

Browse by Faculty

Economic and Management Sciences

Education

Health Sciences

The Humanities

Law

Natural and Agricultural Sciences

Theology and Religion

Explore other UFS collections, archival material, museum and KovsieJournals

Electronic Theses & Dissertations ›Private Book Collection ›National Museum ›Archive for Contemporary Affairs ›KovsieJournals ›

Recent Submissions

Access status: Open Access ,
Long-term Life Insurance valuations meet deep learning
(University of the Free State, 2025) Blomerus, Jan Marthinus; Ring, A.
Traditional actuarial methods for valuing insurance portfolios, while established, are often time consuming, complex, and prone to manual error. This study investigates the potential of machine learning techniques to enhance and streamline these traditional methods, offering improved efficiency and accuracy. Utilising a dataset from a commercial European life insurance company, this research designs and implements a deep neural network to predict policy reserve values. A combination of actuarial pricing and valuation bases is employed to prepare the data for training and evaluation, focusing on developing models capable of accurately predicting expected present values. The results demonstrate that machine learning models can effectively predict policy reserve values, providing valuations comparable to those obtained through traditional actuarial methods. Notably, the developed ”Midway” model consistently predicts accurate and efficient reserve estimates. These models demonstrate an ability to capture complex relationships between inputs and policy reserve values, even with combinations of previously unseen data within valid ranges. This research has significant implications for the insurance industry, offering the potential for improved efficiency, enhanced risk management, and more informed decision-making. The ability to rapidly and accurately value large policy portfolios can lead to improved pricing strategies and investment decisions. Furthermore, machine learning techniques can reduce the time and resources required for traditional valuations and provide an independent check on accuracy for auditors and regulators. Beyond its practical applications, this study contributes to the machine learning community by demonstrating a novel combination of methods for fitting supervised regression models and establishing a framework for data preparation, model training, and validation. This work provides valuable guidance for future research in applying machine learning to diverse categories of life insurance products. In conclusion, this study provides a compelling proof-of-concept for leveraging machine learning techniques to value a commercial book of life insurance policies, highlighting the potential for significant improvements in risk management within the insurance industry.
Access status: Open Access ,
An evaluation of the use of high-resolution RGB and multispectral imaging as a potential tool for drill-core logging of the rocks of the Bushveld Complex
(University of the Free State, 2025) Mabogo, Nomonde Shantel Tshiwela; Roelofse, Frederick; Clark, Martin
Core recovery drilling remains a fundamental component of mineral exploration in the Bushveld Complex of South Africa, which hosts the world’s largest platinum-group element reserves within the world’s largest layered intrusion, the mafic to ultramafic Rustenburg Layered Suite. Manual logging of extensive drill-core intervals is time-consuming and inherently subjective, often limiting the precision of lithological and mineralogical interpretations. This study assesses spectroscopic drill-core logging techniques, such as RGB to-grayscale and visible to near-infrared (VNIR) multispectral imaging (400-842 nm), applied to the Rustenburg Layered Suite drill-core as potential tools for cost-effective drill-core logging of the major lithologies and minerals of the Rustenburg Layered Suite. High-resolution RGB images were converted to grayscale using the luminosity method, and grayscale digital number (0-255) statistics, median and skewness were analysed for lithological classification. The grayscale approach distinguishes chromitite and anorthosite through contrasting histogram skewness, with median-based validation correctly classifying 3% of anorthosite samples and 0% of chromitite, the latter is due to limited representation in the training dataset and that training and validation chromitite samples differ in the proportion of dark and lighter crystals. In contrast, gabbronorite and pyroxenite proved more challenging to differentiate, reflecting their broad compositional ranges and overlapping plagioclase pyroxene assemblages, with 19% gabbronorite validation sites and 44% pyroxenite validation sites being either a gabbronorite or pyroxenite. Additionally, the weighting of the green channel in the luminosity-based grayscale conversion can shift darker pixels toward higher grayscale values, potentially leading to the misclassification of dark lithologies as lighter-coloured lithologies. Thus, this approach proved less effective for reliable lithological classification across the dataset. VNIR multispectral imaging (475-842 nm) enables classification of major minerals such as pyroxene, plagioclase, and chromite, with spectral library correlations improving accuracy by identifying minerals prone to misclassification. The results show that 40% of pyroxenite samples contained over 90% pyroxene, 12% of anorthosite samples contained over 90% plagioclase, 71-75% of gabbronorite samples exhibited mixed plagioclase-pyroxene proportions within the 10-90% compositional range, and all chromitite samples contained more than 45% chromite. Results also reflect expected mineralogical trends, with mafic rocks showing higher mafic mineral content and anorthositic rocks dominated by felsic minerals.
Access status: Open Access ,
Masculinity and Fatherhood: A Narrative Study of South African Black Men Living in Mangaung
(University of the Free State, 2024) Velelo, Nontombi Lenah; Rau, Asta; Coetzee, Jan K.
Access status: Open Access ,
The complexities of translator education in Ghana: exploring a human capabilities approach to curriculum design
(University of the Free State, 2024) Dakey, Linda Esinam; Marais, Kobus; González-Davies, Maria
Access status: Open Access ,
Representations of trauma in poetic parallelisms of the Biblical Hebrew Book of Job: a complexity theoretical analysis
(University of the Free State, 2024) Nortier, Hermias Hendrik Sutherland; Naudé, J.A.; Miller-Naudé, C.L.
Access status: Open Access ,
Morphosyntactic aspects of the infinitive construct in Qumran Hebrew
(University of the Free State, 2024) Rampanjato, Fianarana Andriamahatozo; Miller-Naudé, C.L.; Naudé, J.A.
Access status: Open Access ,
Emerging from Crisis: Representation of Childhood Identities From Selected Zimbabwean Short Stories of the Third Chimurenga Era
(University of the Free State, 2024) Mawire, Primrose Rufaro; Ngara, K.M.; Manase, I.
Access status: Open Access ,
Cosmologies of Madness: Exploring Representations of Madness in Contemporary African Literary and Cultural Texts
(University of the Free State, 2024) Fingson, Kundai Watson; Ngara, Kudzayi M.; Nyambi, Oliver
Access status: Open Access ,
Imaging prosperity in African Christianity: a practical theological case study of urban congregations in Lusaka, Zambia
(University of the Free State, 2024) Banda, Daniel Sandford; Van der Watt, Gideon
Access status: Open Access ,
Narratiewe pastorale spel en gemarginaliseerde gemeenskappe
(University of the Free State, 2024) Van Schalkwyk, Hester Helena; Van den Berg, Jan-Albert