Department of Computational Sciences

Permanent URI for this community

http://hdl.handle.net/20.500.11837/158

The Department of Computational Sciences collection provides access to the scholarly and research outputs of staff and students. This collection includes theses and dissertations, research articles, conference papers, and examination question papers. The Department brings together the disciplines of Computer Science, Pure and Applied Mathematics, and Statistics, offering a strong foundation in analytical, computational, and problem-solving skills. Through its focus on both theoretical and applied approaches, the Department equips students to contribute to advances in technology, data analysis, mathematical modelling, and scientific research. By preserving and disseminating these resources, the collection supports teaching, learning, and research while fostering innovation, critical inquiry, and capacity-building within the University and the broader society.

Browse

Now showing 1 - 20 of 109

A Comparison of Open Source Native XML Database Products
(University of Fort Hare, 2005-12) Mabanza, Ntima
This introductory chapter discusses the motivation for studying XML databases, including some questions that provide direction for the work. It also briefly introduces previous research, the scope of the thesis and also the research methodology. Chapter one concludes by giving the organization of the thesis.
A framework for evaluating the reliability of health monitoring technologies that are based on ambient intelligence
(University of Fort Hare, 2024-12) Scott, Mfundo Shakes; Jere, N.R.; Sibanda, K.
The advancement of health monitoring technologies rooted in Ambient Intelligence (AmI) has created innovative solutions for improving healthcare delivery and supporting independent living. However, a critical challenge persists because of the absence of a standardised, integrated framework to evaluate the reliability of these systems. This gap is particularly significant in resource-constrained environments, where operational uncertainties and infrastructure limitations exacerbate reliability issues. To address this problem, this study proposes a comprehensive framework for evaluating the reliability of AmI-based health monitoring technologies. The framework encompasses key dimensions, including data accuracy, system robustness, reliability and context or environment considerations. It was developed using a combination of simulation-based modelling, reliability block diagrams, and Monte Carlo Markov Chain (MCMC) techniques to systematically quantify and analyse reliability metrics. These methods were applied to case studies such as continuous glucose monitoring and heart rate monitoring in elderly care to validate the framework's practical relevance. Results from these applications demonstrate the framework's efficacy in identifying and addressing reliability challenges. The findings highlight its ability to facilitate systematic evaluation, enhance fault tolerance, and ensure adaptability across diverse operational contexts. The presented framework contributes significantly to various stakeholders as it offers researchers a structured methodology for reliability assessment, provides healthcare practitioners with tools to ensure dependable patient monitoring, and guides policymakers in the adoption of reliable health technologies. By bridging the gap between technological innovation and practical application, this research advances the field of AmI-based health monitoring, fostering improved healthcare outcomes and greater trust in intelligent systems. Software developers, researchers, Health Monitoring devices manufacturers, Internet of Things (IOT) experts and patients on chronic medication who require regular monitoring, and the health sector are the targeted beneficiaries of the framework. Using two case studies which are Continuous Glucose Monitoring and Heart Rate Monitoring in elderly care the main limitations was the effectiveness of reliability in addressing these issues. This lays the groundwork for further refinement and application of the framework in broader contexts. Thus, future research ought to focus on addressing operational challenges, improving explainability in AI models, and discovering innovative applications of AI in healthcare monitoring systems.
A Kmeans-LSTM hybrid model for optimising spectrum sensing in wireless networks
(University of Fort Hare, 2024-03) Tamuka, Nyashadzashe Everson; Sibanda, K.
Due to the increasing demand for the wireless spectrum, cognitive radio (CR) technology has gained attention. Spectrum sensing, an essential task of CR, involves detecting vacant frequency bands for unlicensed users when licensed users are not fully utilising them. Traditional techniques like energy detection and matched-filter detection have limitations, including poor performance at low signal-to-noise ratios (SNR) and the need for prior knowledge of licensed user signal characteristics. Furthermore, most machine learning (ML) and hybrid techniques rely on simulated datasets, lacking real-world validation. This research proposed a novel hybrid Kmeans-LSTM model, robust at low SNR and independent of prior licensed user information. The model was validated with real-world spectrum datasets collected using an RTL-SDR dongle, unlike most of previous studies that depend on simulated datasets. Guided by design-science research methodology, the study involved iterative training, testing, and evaluation. The spectrum data was collected and labelled using K-means, followed by training the LSTM algorithm. Performance metrics, including accuracy, detection probability, false-alarm probability, precisionrecall curves, and ROC curves, showed that the Kmeans-LSTM model outperformed other models such as Support Vector Machine, Artificial Neural Network, and Random Forest. The real-world dataset provided a realistic assessment of the model’s performance, confirming its robustness and potential for optimizing spectrum sensing in wireless networks. This study demonstrated that the Kmeans-LSTM model can optimise spectrum sensing at low Signal-Noise-Ratio conditions without simulated and labelled spectrum datasets.
Advanced Electrodynamics: PHY 222, Supplementary Examinations November 2025
(University of Fort Hare, 2025-11) Mukumba, P.; Kwinana, P.
Advanced Programming in C++: CSC 211, Degree Examinations June 2023
(University of Fort Hare, 2023-06) Ngwenya, S.; Nomnga, P.
An analysis of the correlation beween packet loss and network delay on the perfomance of congested networks and their impact: case study University of Fort Hare
(University of Fort Hare, 2013) Lutshete, Sizwe; Sibanda, K
In this paper we study packet delay and loss rate at the University of Fort Hare network. The focus of this paper is to evaluate the information derived from a multipoint measurement of, University of Fort Hare network which will be collected for a duration of three Months during June 2011 to August 2011 at the TSC uplink and Ethernet hubs outside and inside relative to the Internet firewall host. The specific value of this data set lies in the end to end instrumentation of all devices operating at the packet level, combined with the duration of observation. We will provide measures for the normal day−to−day operation of the University of fort hare network both at off-peak and during peak hours. We expect to show the impact of delay and loss rate at the University of Fort Hare network. The data set will include a number of areas, where service quality (delay and packet loss) is extreme, moderate, good and we will examine the causes and impacts on network users.
Building a semantic web-based e-health component for a multi-purpose communication centre
(University of Fort Hare, 2010) Hlungulu, Bulumko
Rural communities have limited access to health information which is made available on the internet. This is due to poor infrastructure (i.e., lack of clinics or Internet access) and that gives them problems in accessing information within the domain of health. The availability of Information and Communication Technologies (ICTs) in a rural community can provide the community with a number of beneficial solutions to their problems as they maximize the potential of knowledge sharing and delivery. This research seeks to make use of ICTs deployed in the community of Dwesa, in order to contribute to improving the health standards of the community. It seeks to accomplish this by carrying out an investigation and literature review with the aim of understanding health knowledge sharing dynamics in the context of marginalized communities. The knowledge acquired will then be used in the development and implementation of a semantic web-based e-Health portal as part of the Siyakhula Living Lab (SLL) project. This portal will share and deliver western medical knowledge, traditional knowledge and indigenous knowledge. This research seeks to make use of a combination of Free and/or Open Sources Software in developing the portal to make it affordable to the community.
Building a semantic web-based e-health component for a multipurpose communication centre
(University of Fort Hare, 2010) Hlungulu, Bulumko; Thinyane, M
Rural communities have limited access to health information which is made available on the internet. This is due to poor infrastructure (i.e., lack of clinics or Internet access) and that gives them problems in accessing information within the domain of health. The availability of Information and Communication Technologies (ICTs) in a rural community can provide the community with a number of beneficial solutions to their problems as they maximize the potential of knowledge sharing and delivery. This research seeks to make use of ICTs deployed in the community of Dwesa, in order to contribute to improving the health standards of the community. It seeks to accomplish this by carrying out an investigation and literature review with the aim of understanding health knowledge sharing dynamics in the context of marginalized communities. The knowledge acquired will then be used in the development and implementation of a semantic web-based e-Health portal as part of the Siyakhula Living Lab (SLL) project. This portal will share and deliver western medical knowledge, traditional knowledge and indigenous knowledge. This research seeks to make use of a combination of Free and/or Open Sources Software in developing the portal to make it affordable to the community.
The classification performance of Bayesian networks classifiers : a case study of detecting Denial of Service (DoS) attacks in cloud computing environments
(University of Fort Hare, 2016) Moyo, Lindani
In this research we propose a Bayesian networks approach as a promissory classification technique for detecting malicious traffic due to Denial of Service (DoS) attacks. Bayesian networks have been applied in numerous fields fraught with uncertainty and they have been proved to be successful. They have excelled tremendously in classification tasks i.e. text analysis, medical diagnoses and environmental modeling and management. The detection of DoS attacks has received tremendous attention in the field of network security. DoS attacks have proved to be detrimental and are the bane of cloud computing environments. Large business enterprises have been/or are still unwilling to outsource their businesses to the cloud due to the intrusive tendencies that the cloud platforms are prone too. To make use of Bayesian networks it is imperative to understand the ―ecosystem‖ of factors that are external to modeling the Bayesian algorithm itself. Understanding these factors have proven to result in comparable improvement in classification performance beyond the augmentation of the existing algorithms. Literature provides discussions pertaining to the factors that impact the classification capability, however it was noticed that the effects of the factors are not universal, they tend to be unique for each domain problem. This study investigates the effects of modeling parameters on the classification performance of Bayesian network classifiers in detecting DoS attacks in cloud platforms. We analyzed how structural complexity, training sample size, the choice of discretization method and lastly the score function both individually and collectively impact the performance of classifying between normal and DoS attacks on the cloud. To study the aforementioned factors, we conducted a series of experiments in detecting live DoS attacks launched against a deployed cloud and thereafter examined the classification performance in terms of accuracy of different classes of Bayesian networks. NSL-KDD dataset was used as our training set. We used ownCloud software to deploy our cloud platform. To launch DoS attacks, we used hping3 hacker friendly utility. A live packet capture was used as our test set. WEKA version 3.7.12 was used for our experiments. Our results show that the progression in model complexity improves the classification performance. This is attributed to the increase in the number of attribute correlations. Also the size of the training sample size proved to improve classification ability. Our findings noted that the choice of discretization algorithm does matter in the quest for optimal classification performance. Furthermore, our results indicate that the choice of scoring function does not affect the classification performance of Bayesian networks. Conclusions drawn from this research are prescriptive particularly for a novice machine learning researcher with valuable recommendations that ensure optimal classification performance of Bayesian networks classifiers.
The classification performance of Bayesian Networks Classifiers: a case study of detecting Denial of Service (DoS) attacks in cloud computing environments
(University of Fort Hare, 2015) Moyo, Lindani; Sibanda, Khulumani
In this research we propose a Bayesian networks approach as a promissory classification technique for detecting malicious traffic due to Denial of Service (DoS) attacks. Bayesian networks have been applied in numerous fields fraught with uncertainty and they have been proved to be successful. They have excelled tremendously in classification tasks i.e. text analysis, medical diagnoses and environmental modeling and management. The detection of DoS attacks has received tremendous attention in the field of network security. DoS attacks have proved to be detrimental and are the bane of cloud computing environments. Large business enterprises have been/or are still unwilling to outsource their businesses to the cloud due to the intrusive tendencies that the cloud platforms are prone too. To make use of Bayesian networks it is imperative to understand the ―ecosystem‖ of factors that are external to modeling the Bayesian algorithm itself. Understanding these factors have proven to result in comparable improvement in classification performance beyond the augmentation of the existing algorithms. Literature provides discussions pertaining to the factors that impact the classification capability, however it was noticed that the effects of the factors are not universal, they tend to be unique for each domain problem. This study investigates the effects of modeling parameters on the classification performance of Bayesian network classifiers in detecting DoS attacks in cloud platforms. We analyzed how structural complexity, training sample size, the choice of discretization method and lastly the score function both individually and collectively impact the performance of classifying between normal and DoS attacks on the cloud. To study the aforementioned factors, we conducted a series of experiments in detecting live DoS attacks launched against a deployed cloud and thereafter examined the classification performance in terms of accuracy of different classes of Bayesian networks. NSL-KDD dataset was used as our training set. We used ownCloud software to deploy our cloud platform. To launch DoS attacks, we used hping3 hacker friendly utility. A live packet capture was used as our test set. WEKA version 3.7.12 was used for our experiments. Our results show that the progression in model complexity improves the classification performance. This is attributed to the increase in the number of attribute correlations. Also the size of the training sample size proved to improve classification ability. Our findings noted that the choice of discretization algorithm does matter in the quest for optimal classification performance. Furthermore, our results indicate that the choice of scoring function does not affect the classification performance of Bayesian networks. Conclusions drawn from this research are prescriptive particularly for a novice machine learning researcher with valuable recommendations that ensure optimal classification performance of Bayesian networks classifiers.
A comparison of open source object-oriented database products
(University of Fort Hare, 2009) Khayundi, Peter; Chadwick, J
Object oriented databases have been gaining popularity over the years. Their ease of use and the advantages that they offer over relational databases have made them a popular choice amongst database administrators. Their use in previous years was restricted to business and administrative applications, but improvements in technology and the emergence of new, data-intensive applications has led to the increase in the use of object databases. This study investigates four Open Source object-oriented databases on their ability to carry out the standard database operations of storing, querying, updating and deleting database objects. Each of these databases will be timed in order to measure which is capable of performing a particular function faster than the other.
Complex Analysis: MAT 322, Degree Examinations October/November 2024
(University of Fort Hare, 2024-10) Prins, A.L.; Nkonkobe, S.
Complex Analysis: MAT 322, Degree Examinations October/November 2025
(University of Fort Hare, 2025-10) Prins, A.L.; Seretlo, T.T.
Complex Analysis: MAT 322, Supplementary Examinations January 2019
(University of Fort Hare, 2019-01) Ndiweni, O.; Murali, V.
Complex Analysis: MAT 322, Supplementary Examinations November 2024
(University of Fort Hare, 2024-11) Prins, A.L.; Nkonkobe, S.
This supplementary examination paper for MAT 322, "Complex Analysis," is a 3-hour assessment worth a maximum of 100 marks. Students are required to answer five questions. The paper covers various advanced topics in complex analysis. The first section of the exam consists of multiple-choice questions testing foundational concepts such as the image of a line under a linear transformation, limits of complex functions, differentiability of complex functions, convergence of sequences, and regions of convergence for complex series. It also includes finding values of complex numbers for which an exponential equation holds. The second section focuses on properties of complex functions, requiring students to prove whether a given function is entire, define elementary functions like the complex cosine function and prove its differentiability and derivative, and demonstrate that a specific function is harmonic, subsequently finding its harmonic conjugate. A portion of the exam delves into contour integration and Cauchy's theorems. Students are required to evaluate contour integrals using the Cauchy Integral Formula and the Cauchy-Goursat theorem, and apply the Cauchy Integral Formula for derivatives to compute specific integrals. Further sections cover Taylor and Laurent series expansions and singularity classification. This includes computing the Taylor series for a given function centered at a specific point and finding its interval of convergence. Students must also compute the Laurent series representation for a function in a given domain and classify the singularities of another function. The examination also includes a section on residue theory, requiring students to state the Cauchy Residue Theorem, find residues of functions at given poles, and use the theorem to evaluate contour integrals.
Computational Methods: MNU 122, Degree Examinations October/November 2025
(University of Fort Hare, 2025-10) Katywa, N.; Mahlasela, Z.
Computational Methods: MNU 122, Supplementary Examinations January 2019
(University of Fort Hare, 2019-01) Nxala, B.E.; Mahlasela, Z.
Computational Methods: MNU 122, Supplementary Examinations November 2024
(University of Fort Hare, 2024-11) Mtakazi, S.; Katywa, N.
This supplementary examination paper for MNU 122, "Computational Methods," is a 3-hour assessment worth 100 marks. Students are required to answer all questions, which cover various topics in linear algebra, numerical methods, and mathematical modeling. Question 1 (18 marks) focuses on linear equations and matrix properties. It requires solving a system of linear equations using the Gaussian Elimination Method and proving a property related to matrix multiplication and addition for given matrices. Question 2 (35 marks) delves into linearizing non-linear equations and applying the least squares method. It asks students to express various non-linear equations in linear form. Subsequently, it presents a practical problem involving torque and diameter, where students must determine the best values for constants in a power law relationship using the least squares method from given data. Another part of this question involves finding the best values for constants in a reciprocal relationship from a given set of data. Question 3 (35 marks) covers differential equations and their applications. It defines a differential equation and asks students to determine the truthfulness of statements regarding the homogeneity, exactness, and linearity of given differential equations, correcting false statements. It then presents a real-world scenario involving the growth of supermarkets using a computerized checkout system, described by an initial value problem. Students must find the number of supermarkets at any time t and compute the number of supermarkets at a specific time (t=10). The question concludes with solving various types of differential equations (homogeneous, non-homogeneous, and Cauchy-Euler equations). Question 4 (12 marks) focuses on Laplace transforms. It requires showing properties of Laplace transforms, specifically for scaling and exponential functions, and finding the Laplace transform of a trigonometric function. It also involves working with a piecewise function and finding its Laplace transform. The examination aims to comprehensively assess students' theoretical understanding and practical application of computational methods in solving mathematical and real-world problems.
Cultural and linguistic localization of the virtual shop owner interfaces of e commerce platforms for rural development
(University of Fort Hare, 2009) Dyakalashe, Siyabulela; Muyingi, H N; Terzoli, A
The introduction of Information and Communication Technologies (ICTs) for rural development in rural marginalized societies is vastly growing. However, the success of developing and deploying ICT related services is still in question as influential factors such as adaptability, scalability, sustainability, and usability have great effect on the rate of growth of ICTs in rural environments. The problem is that these ICT services should be maintained and sustained by the targeted communities. The main cause for rural marginalization is the fact that some communities situated in rural settings are educationally challenged and computer illiterate or semiliterate in comparison with urban communities. An ICT for development (ICT4D) intervention in the form of an e-Commerce platform that targets the social and economic growth of rural marginalized communities has been developed and field tested at Dwesa, a rural community located on the Wild Coast of the former homeland of Transkei in the Eastern Cape Province. The e-Commerce platform is known as “buy at Dwesa” and can be visited at this URL, http://www.dwesa.com. The aim of the e-Commerce platform is to motivate small entrepreneurs in rural areas to market their products and themselves to the global market as they lack the skills and resources for marketing their art and crafts. Virtual stores are created for a small group of entrepreneurs who will maintain and sustain the stores on their own. These entrepreneurs are often elderly women with limited education and little to no computer literacy - meaning that sustaining the stores may prove difficult for them. In this research we discuss the re-design and re-development of the virtual shop-owner interfaces of the e-Commerce platform to make them more culturally and linguistically localized. The virtual shops allow shop-owners to upload their artifacts to advertise and sell on the customer’s end of the e-Commerce platform. For multilingual and multicultural communities, adoption of the software interfaces to the user’s cultural and linguistic needs and modes of expression is important as failure to do so may reduce the level of benefits of e-Commerce initiatives.
Cultural and Linguistic Localization of the Virtual Shop-Owner Interfaces of ECommerce Platforms for Rural Development
(University of Fort Hare, 2009-11) Dyakalashe, Siyabulela
The introduction of Information and Communication Technologies (ICTs) for rural development in rural marginalized societies is vastly growing. However, the success of developing and deploying ICT related services is still in question as influential factors such as adaptability, scalability, sustainability, and usability have great effect on the rate of growth of ICTs in rural environments. The problem is that these ICT services should be maintained and sustained by the targeted communities. The main cause for rural marginalization is the fact that some communities situated in rural settings are educationally challenged and computer illiterate or semiliterate in comparison with urban communities.

Browse

Browsing Department of Computational Sciences by Title

Results Per Page

Sort Options