Home /  Proceedings

Proceedings

All full papers submissions to SAICSIT 2024 were double-blind peer-reviewed by an international panel of experts.  Accepted papers are published in Springer CCIS Series - Communications in Computer and Information Science.  This publication can be accessed via the following link:

https://link.springer.com/book/10.1007/978-3-031-64881-6​

Access to this publication will be free for conference participants for three weeks starting from the first day of the conference.

High quality papers not accepted for the CCIS series are published in the SAICSIT 2024 online proceedings.  This publication can be accessed via the following link:

SAICSIT 2024 Online Proceedings

Work-in-Progress Papers were presented at the Postgraduate Symposium and are included in the SAICSIT 2024 Work-in-Progress Proceedings.  This publication can be accessed via the following link: Work-in-Progress Papers



 

Springer Papers with Abstracts

Carr, L., Chavula, J. (2024). Deep Learning Classification for Encrypted Botnet Traffic: Optimising Model Performance and Resource Utilisation. In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer. View more...

Abstract
Detection of malicious traffic on a network is critical to ensuring the safety and security of internet systems. Classical approaches to this task increasingly struggle with modern networking procedures, like encryption. Deep learning (DL) offers an alternative approach to traffic classification problems. We address two major problem classes: (1) botnet detection and (2) botnet family classification. For each problem, we explore five implementations of DL architectures: a multi-layer perceptron (MLP), shallow and deep convolutional neural network (CNN v1 and CNN v2), an autoencoder (AE) and an autoencoder + convolutional neural network (AE+CNN). Our evaluation of models for each respective problem class is based on the classification performance and computational requirements of each model. We further investigate the effect of training the models on an input with a reduced feature space, where we evaluate the impact this has in terms of a trade-off between computational and classification performance. For botnet detection, we find that all models attain good (≥0.979 accuracy) classification performance on a normal testing set; however, this performance drops fairly substantially when evaluated on a set of unknown botnet families. Furthermore, we observed a clear trend between increased feature space and memory utilisation, while finding no evidence of a trend between inference time and feature space. For botnet classification, we found that models which implement CNN architectures outperform others by a substantial margin (≈6 percentage points). We observe the same trend between feature space and memory utilisation, and absence of apparent relationship between feature space and inference time.

 

Chalwe, C., Chanda, C., Muzyamba, L., Mwape, J., Phiri, L. (2024). Quantitative Analysis of Zambian Wikipedia Contributions: Assessing Awareness, Willingness, Motivation, and the Impact of Gamified Leaderboards and Badges. In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer  View more...

Abstract
Wikipedia is a widely recognized and valuable source of information, However, it encounters persistent challenges in attracting and retaining active contributors. It is recorded that only 10 people from Zambia contribute and create content on wikipedia in the month of may 2023. While a large number consumes Wikipedia content, there is a noticeably low number of Wikipedians that contribute content on and about Zambia. This paper presents a Facebook plugin, WikiMotivate, aimed at motivating Zambian Wikipedians to update pre-existing content, add new entries, and share their natural expertise. WikiMotivate was implemented as a Facebook plugin that utilizes leaderboard and badge gamification features to encourage and incentivize active Wikipedia content contribution. Using a mixed-methods approach, historical Wikipedia edit histories were used to quantify content contributed by Zambian Wikipedians. In addition, user surveys were conducted to determine relative levels of awareness about Wikipedia, willigness to contribute to contribute content on Wikipedia and perceived motivating factors that affect content contribution on Wikipedia. Furthermore, a Facebook plugin, WikiMotivate, was implemented in order to be used an a service for motivating potential Zambian Wikipedians. Finally, in order to determine the most effective approach, a comparative analysis of leader-boards and badges was conducted with nine (9) expert evaluaters. The results clearly indicate that a significant proportion of Wikipedia content on and about Zambia is authored by Wikipedians from outside Zambia, with only 11% of the contributors, out of the 224, originating from Zambia. In addition, study participants were largely unaware of the various editing practices on Wikipedia; interestingly enough, most participants expressed their willingness to contribute content if trained. In terms of motivating factors, “Information Seeking and Educational Fullfilment” was as the key motivating factor. The Facebook plugin implemented suggests that incorporating leaderboards and badges is a more effective approach to motivating contributions to Wikipedia. This study provides useful insight into the landscape of Wikipedia content contribution in the Global South.

 

Herbst, C., Jeantet, L., Dufourq, E. (2024). Empirical Evaluation of Variational Autoencoders and Denoising Diffusion Models for Data Augmentation in Bioacoustics Classification. In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer. View more...

Abstract
One major challenge in supervised deep learning is the need for large training datasets to achieve satisfactory generalisation performance. Acquiring audio recordings of endangered animals compounds this issue due to high costs, logistical constraints, and the rarity of the species in question. Typically, bioacoustic datasets have imbalanced class distributions, further complicating model training with limited examples for some rare species. To overcome this, our study proposes the evaluation of generative models for audio augmentation. Generative models, such as Variational Autoencoders (VAEs) and Denoising Diffusion Probabilistic Models (DDPMs), offer the ability to create synthetic data after training on existing datasets. We assess the effectiveness of VAEs and DDPMs in augmenting a bioacoustics dataset, which includes vocalisations of the world’s rarest primate, the Hainan gibbon. We assess the generated synthetic data through visual inspection and by computing the Kernel Inception Distance, to compare the distribution of the generated dataset to the training set. Furthermore, we investigate the efficacy of using the generated dataset to train a deep learning classifier to identify the Hainan gibbon calls. We vary the size of the training datasets and compare the classification performance across four scenarios: no augmentation, augmentation with VAEs, augmentation with DDPMs, and standard bioacoustics augmentation methods. Our study is the first to show that standard audio augmentation methods are as effective as newer generative approaches commonly used in computer vision. Considering the high computational costs of VAEs and DDPMs, this emphasises the suitability of simpler techniques for building deep learning classifiers on bioacoustic datasets.

 

Kandjimi, H., Suleman, H. (2024). Investigating Markov Model Accuracy in Representing Student Programming Behaviours. In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer. View more...

Abstract
Problem-solving skills are an integral component within the computer science field. Due to the diversity brought about by students following different learning and programming behaviours, it is challenging to track and identify when students get overwhelmed while writing programs. When students are overwhelmed, they are unable to complete learning objectives on time and follow prescribed pathways, depriving them of the opportunity to learn new concepts. In this paper, we developed and evaluated the quality of Markov models that encode student programming behaviours based on the evolution of source code submissions during formative practical assignments. In doing so, we use Abstract Syntax Trees (ASTs) extracted from the source code, which are used for clustering similar submissions and tracking students’ progressive approaches within the Markov models. An approach based on MinHashLSH is presented that works on AST nodes as input to emphasise structural similarity and related programming approaches. As such, the effectiveness of the Modified MinHashLSH approach is based on the clusters that make up the Markov model. The research result shows that we can successfully create a highquality model based on previous data. This model result could be used to inform the development of learning interventions that would move students from their stuck states.

 

Mamba, E.B., Levitt, S.P. (2024). A Longitudinal Study on the Effect of Patches on Code Coverage and Software System Maintainability. In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer View more...

Abstract
Contemporary software development often involves the use of source control repositories which are hosted online and making incremental patches (commits) throughout the development process. Online hosting services facilitate the use of build pipelines and the integration of code coverage services into these pipelines. However, existing research into how incremental patches to software systems affect code coverage has not fully taken advantage of the data that is made available by these coverage services. This paper presents a partial replication of two previous studies on patch coverage, analysing over 50,000 builds from 46 projects obtained from two popular coverage services, Codecov and Coveralls. Data quality issues, such as missing commits, duplicate builds from Cron Jobs, and sudden coverage drops, were identified and addressed, highlighting the need for rigorous data cleaning process when mining data from coverage services. Results indicate that patches are generally either fully covered or entirely uncovered, with a majority achieving full coverage, suggesting that very seldom do engineers opt for partial coverage. There is a weak correlation (correlation coefficient: 0.23) between patch coverage and system coverage, indicating that patch coverage alone cannot be used to predict system coverage evolution. Furthermore, patch testing does not enhance patch maintainability.

 

Martin, R.H.J., Visser, R., Dunaiski, M. (2024). Is Transformer-Based Attention Agnostic of the Pretraining Language and Task?. In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer View more...

Abstract
Since the introduction of the Transformer by Vaswani et al. in 2017, the attention mechanism has been used in multiple state-ofthe- art large language models (LLMs), such as BERT, ELECTRA, and various GPT versions. Due to the complexity and the large size of LLMs and deep neural networks in general, intelligible explanations for specific model outputs can be difficult to formulate. However, mechanistic interpretability research aims to make this problem more tractable. In this paper, we show that models with different training objectives—namely, masked language modelling and replaced token detection—have similar internal patterns of attention, even when trained for different languages, in our case English, Afrikaans, Xhosa, and Zulu. This result suggests that, on a high level, the learnt role of attention is language-agnostic.

 

Mertens, P.J., Ngxande, M. (2024). Age-Related Face Recognition Using Siamese Networks and Vision Transformers. In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer View more...

Abstract
Face recognition plays a crucial role in various applications, ranging from security to personal convenience. Recent advancements have emphasized the importance of recognizing individuals based on age-related facial features within this domain. This paper presents a comprehensive evaluation of two deep learning architectures for age-based face recognition: Siamese Convolutional Networks (SCNs) and Vision Transformers (ViTs). Convolutional Neural Networks (CNNs), which are critical in modern face recognition, serve as the backbone for Siamese Convolutional Networks (SCNs). SCNs are specifically designed to detect similarities between input pairs by emphasising local features crucial for age-related distinctions. In contrast, ViTs, initially developed for natural language processing, have demonstrated promising performance in image recognition, showcasing their aptitude for capturing global image context. This work investigates the performance of these distinct architectures in discerning age-related variations within facial data features. Performance comparisons were conducted on three established SCN models and two ViT architectures. The results revealed that the optimal SCNs primarily focused on the mouth, nose, and eye regions, indicating their reliance on local features for age estimation. Interestingly, the ViT models achieved superior performance despite lacking explicit feature localization. This suggests that a holistic understanding of the facial context may be more effective than focusing solely on isolated features for age-based recognition.

 

Thikho, M., Mokwena, S.N. (2024). Sarcasm Detection in Political Speeches Using Recurrent Neural Networks.  In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer View more...

Abstract
This study investigated the effectiveness of recurrent neural networks (RNN) for the detection of sarcasm in the challenging domain of political speech. Given the inherent nuance of sarcasm, it can be difficult. This study compared three RNN architectures (SimpleRNN, LSTM, and GRU) and demonstrated that ensemble learning techniques (stacking andweighted averaging) further improved accuracy. Pre-trained word embeddings (GloVe) were used to capture semantic cues that signal sarcasm. These embeddings were incorporated by replacingwords with their corresponding vector representations. The model was evaluated using standard metrics (accuracy, precision, recall, F1 score). The results showed that the ensembles outperformed individual RNNs, achieving a peak accuracy of 95% and an F1 score of 95% for both sarcastic and nosarcastic classes. IndividualRNNs achieved an accuracy of 91%, highlighting the clear benefit of ensemble learning. This improvement suggests that ensembles effectively combine the strengths of different models, leading to more robust and generalisable sarcasm detection in political speech. Furthermore, this research paves the way for the application of similar techniques to sentiment analysis tasks in other complex domains.

 

Mulder, S., Du Plessis, M.C. (2024). Unsupervised State Encoding in Video Sequences Using β-Variational Autoencoders. In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer View more...

Abstract
Monitoring and providing feedback on the execution of sequential tasks is common in various domains, such as industrial processes for quality control, automated supervision for skill acquisition and even surgical procedures. This research explores the use of a Disentangled β-Variational Autoencoder (β-VAE) to encode different states in video data depicting a sequence of actions being performed on a series of objects without explicit labels. We trained a Disentangled β-VAE on video data of a sequence being performed and evaluated its ability to distinguish between states using visualisations based on similarity metrics. The evaluation was performed using a set of sequences specifically designed to establish the validity and limits of β-VAE’s encoding of the states. These sequences included both unseen sequences which were similar to the training data, as well as out-of-distribution sequences which deviate from those seen in training. The results demonstrate that the β- VAE successfully learned to encode distinct states within the sequence, as evidenced by the visualisations. It is shown that β-VAE is capable of detecting states within a sequence. Furthermore, it is demonstrated that these learnt states inherently also have learnt dependencies and relationships regarding the sequence in which they are performed. These findings lay the foundation for the development of an overarching algorithm that monitors a sequence in progress and provides feedback when deviations from the expected sequence occur.

 

Shawa, A., Chileshe, E., Mwaba, B., Mwanza, J., Sikazwe, W., Zulu, E.O., Phiri, L. (2024). User Centred Design and Implementation of Useful Picture Archiving and Communication Systems for Effective Radiological Workflows in Public Health Facilities in Zambia. In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer View more...

Abstract
Radiological workflows in public health facilities in The Republic of Zambia are performed using manual processes. With a broad spectrum of stakeholders—physicians, radiographers and radiologists involved in radiological workflows, the efficiency of health service provision is drastically reduced, subsequently compromising clinical care. While there are a number of software platforms that are used in radiological workflows, Picture Archiving and Communication System platforms are important as they are primarily used to store, manage and facilitate access to Medical Images. This paper outlines the user-centred design and implementation of a Picture Archiving and Communication System for storing, managing, and facilitating access to medical images in public health facilities in Zambia, in order to demonstrate the feasibility of automating manual medical imaging workflows in public health facilities. Semi-structured interviews were conducted with two (2) radiologists and four (4) radiographers in order to understand medical imaging workflows in public health facilities. A Picture Archiving and Communication System was designed and implemented using Dicoogle as the base platform and, subsequently evaluated—using the TAM 2 questionnaire— in order to assess its perceived usability and usefulness. The interviews conducted provide insight into the extent towards which manual workflows are employed, with Change to Digital Imaging and Communications in Medicine (DICOM) Viewers used as the main technology in the workflow. The implementation and evaluation of the Picture Archiving and Communication System demonstrates the feasibility of implementing these platforms in public health facilities and their potential usefulness, respectively.

 

Ohene-Kwofie, D., Hazelhurst, S. (2024). Single Matrix Block Shift (SMBS) Dense Matrix Multiplication Algorithm. n: South African Computer Science and Information Systems Research Trends. SAICSIT 2024.  Communications in Computer and Information Science, vol 2159. Springer View more...

Abstract
Many scientific and numeric computations rely on matrixmatrix multiplication as a fundamental component of their algorithms. It constitutes the building block in many matrix operations used in numeric solvers and graph theory problems. Several algorithms have been proposed and implemented for matrix-matrix multiplication, especially, for distributed-memory systems, and these have been greatly studied. In particular, the Cannon’s algorithm has been implemented for distributedmemory systems, mostly since the memory needs remain constant and are not influenced by the number of processors employed. The algorithm, however, involves block shifting of both matrices being multiplied. This paper presents a similar block-oriented parallel algorithm for matrixmatrix multiplication on a 2-dimensional processor grid, but with block shifting restricted to only one of the matrices. We refer to this as the Single Matrix Block Shift (SMBS) algorithm. The algorithm, we propose, is a variant of the Cannon’s algorithm on distributed architectures and improves upon the performance complexity of the Cannon and SRUMMA algorithms. We present analytic as well as experimental comparative results of our algorithm with the standard Cannon’s algorithm on 2-dimensional processor grids, showing over 4X performance improvement.

 

Sekokotoana, M., Mhlongo, S., Ade-Ibijola, A. (2024). Automatic Supervision of Online Assessments Using System Process Information and Random Photography.  In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer View more...

Abstract
Recently, online teaching and learning have seen a notable uptrend in adoption, subsequently increasing interest in conducting online assessments. The limitation of remote online assessments lies in the challenge of supervising the individual being assessed. For this reason, many consider human supervision a superior method for maintaining the integrity of assessments. This paper introduces algorithm-driven techniques for the automated supervision of online assessment-takers by analysing system processes on their devices and conducting random photographic monitoring. These techniques, along with their associated algorithms, have been encapsulated into a proof of concept tool. The approach aims to deter assessment-takers from accessing unauthorised files on their devices during assessments and to instil a sense of being monitored. The system is built around two primary components: one that monitors process activity and another that analyses images captured through the assessment-taker’s device webcam. Data collected through these methods are further analysed using facial recognition and additional algorithms to detect behaviours potentially indicative of cheating during the assessment. Initial testing of the proposed tool achieved a 96.3% accuracy rate in image analysis for identifying cheating behaviour. Moreover, university lecturers’ evaluations strongly support the tool’s potential to deter cheating, its effectiveness in detection, and its role in maintaining the integrity of online assessments. Future research is recommended to address the challenges identified with the proof of concept tool, with the objective of enhancing both the accuracy and the overall effectiveness of the proposed techniques.

 

Timm, N., Smith, K. (2024). Iterative Approximation of Nash Equilibrium Strategies for Multi-agent Systems.  In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024.  Communications in Computer and Information Science, vol 2159. Springer View more...

Abstract
We present a technique for approximating Nash equilibrium strategies for multi-agents systems for resource allocation (MRAs). Agents in MRAs seek to maximise the frequency of reaching their resource allocation goals, which can be measured by means of a payoff. A strategy is an ε approximation of a Nash equilibrium if no agent can multiply its pay-off by more than ε via unilateral deviation, where ε is a rational number. For small ε’s approximate Nash equilibria are practically useful and stable strategies since agents will only have a small incentive to change their strategic behaviour. Our technique is based on encoding the strategy synthesis problem in propositional logic with weighted ‘pay-off’ clauses and solving it via weighted maximum satisfiability solving. In our approach we initially synthesise a collectively optimal strategy and determine for each agent the improvement potential, which is the pay-off increase that can be achieved via unilateral deviation. We seek to iteratively reduce the improvement potentials of synthesised strategies: The weights of the ‘pay-off’ clauses associated with agents get adjusted such that agents with a currently high improvement potential will be favoured when solving the weight-adjusted strategy synthesis problem in the subsequent iteration. We show that our approach facilitates the synthesis of ε-equilibrium strategies with small ε’s

 

Tissink, K.D., du Plessis, M.C. (2024). A Comparison of Text Representation Techniques and Encoder-Decoder Implementations in a Deep Neural Network for Converting Natural Language into Formal Logic Formulas. In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer View more...

Abstract
Semantic parsing is the task of extracting a structured machine-interpretable representation from natural language utterance. This representation can be used for various applications such as question answering, information extraction, and dialogue systems. However, semantic parsing is a challenging problem that requires dealing with the ambiguity, variability, and complexity of natural language. This paper investigates neural parsing of natural language (NL) sentences to firstorder logic (FOL) formulas. FOL is a widely used formal language for expressing logical statements and reasoning. FOL formulas can capture the meaning and structure of NL sentences in a precise and unambiguous way. FOL parsing is approached as a sequence-to-sequence mapping task using both long short-term memory (LSTM) and transformer encoder-decoder architectures for character-, subword-, and word-level text tokenisation. These models are trained on NL-FOL datasets with supervised learning and evaluated using various metrics. Previous solutions to neural FOL parsing differ dramatically in training approaches and scale. As such, there is no comprehensive comparison of models for different methods of text representations or encoder-decoder architectures. The main contributions of this paper are: the formation of a complex NL-FOL benchmark that includes algorithmically generated and humanannotated FOL formulas, evaluation of 15 sequence-to-sequence models on the task of neural FOL parsing for different text representations and encoder-decoder architectures, and an in-depth analysis of the strengths and weaknesses of these models.

 

Alabi, O.O., Ajagbe, S.A., Kuti, O., Afe, O.F., Ajiboye, G.O., Adigun, M.O. (2024). Leveraging Environmental Data for Intelligent Traffic Forecasting in Smart Cities. In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer View more...

Abstract
This research revolves around the intersection of environmental data and smart city infrastructure to develop an innovative approach for forecasting traffic patterns. In an era of urbanization and the proliferation of smart cities, managing traffic congestion is a critical challenge. This research explores the utilization of air pollution data, a readily available environmental metric, to intelligently predict traffic patterns and improve urban mobility. The study will delve into the potential correlations between air pollution levels and traffic congestion, considering factors such as vehicular emissions, weather conditions, and geographical attributes. By harnessing the power of big data analytics and machine learning techniques, this research aims to develop a predictive model that leverages real-time air pollution data for traffic forecasting. The K-Nearest Neighbors (KNN) model performs better than all other regression models evaluated in this study, according to our findings. The KNN model considerably lowers the error rate in traffic congestion prediction by more than 28%, according to experimental results.

 

Butale, M., Seymour, L.F. (2024). Understanding Limited Enterprise Systems Benefits in Government: An African Case. In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer View more...

Abstract
Enterprise systems (ES) are increasingly being used in government. Yet the benefits being realized are lower than anticipated and rarely match the level of investment. Post-implementation is when, through usage, benefits are realized and yet it is the ES phase researched the least. Hence there is little understanding of why benefits are limited in government departments. Therefore, this study aimed to answer the question ‘Why are limited ES benefits realized in government?’ The case studies chosen were in the government of an African developing country. Data included semi-structured interviews and secondary data. The explanatory model, inductively derived, reveals no ES benefits management processes in the cases. Low usage and limitations in their ES were predominantly due to the limited IT support post-implementation although limitations with the initial implementation project were raised. The root causes of limited benefits are organizational constraints within the government departments and poor consultant knowledge transfer.

 

Cullen, M., Calitz, A.P., Claassen, N. (2024). Gender Perceptions of Software Developers’ Job Satisfaction. In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer View more...

Abstract
In a competitive market, Information Technology (IT) organisations must retain their most valuable assets, namely the IT employees, specifically software developers. Employee job satisfaction is important for organisations because it influences employee motivation, productivity and commitment to the organisation. The link between gender and job satisfaction has been extensively researched, however limited research has been conducted in the IT industry in South Africa. It is therefore necessary to understand the factors that influence job satisfaction in the IT industry for each gender. Evaluating the career satisfaction of software developers and their intent to remain in their current positions are important for IT organisations. The literature review for this study focused on motivating and hygiene factors that influence job satisfaction and the intention to resign in the IT industry. This research study used a positivistic approach and a quantitative survey. The study established that statistically significant differences regarding gender exist for different factors influencing job satisfaction and the intention to resign at a large South African IT organisation. Gender related statistical differences were found for the factors Career Development, Career Advancement, Leadership, Working Environment – Onsite, Working Environment – Home, Corporate Culture, Company Reputation, Rewards, Purpose, Job Security and Intention to Resign. However, for the factors Job Satisfaction and Organisational Commitment no statistical significant differences were found between male and female software developers. The study is the first study investigating gender differences on job satisfaction and intention to resign in the IT industry in South Africa.

 

Kent, A., Ruhwanya, Z. (2024). Understanding Farmer Perceptions: Impacts on Agricultural IoT Adoption in Western Cape, South Africa. In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer View more...

Abstract
The Internet of Things (IoT) is a transformative technology with applications in various fields, such as medicine, manufacturing and agriculture.Within agriculture, IoT has been shown to improve food security and profitability while saving costs and resources. Further, studies relating to social aspects of IoT, such as user perspectives and adoption, are limited. The Western Cape’s agricultural sector is vulnerable to challenges such as food insecurity and the effects of climate change. Despite this, the adoption of agricultural IoT is uneven within this sector. This study interviewed 17 farmers to gain insights into farmers’ perceptions of agricultural IoT, focusing on how these perceptions influence the adoption of IoT within the Western Cape. Pertinent literature was used to construct a combined framework of potential reasons for and against the adoption of IoT within the Western Cape. This research followed a deductive approach by comparing the combined framework with feedback provided by farmers during the interviews. Farmers within theWestern Cape cited multiple reasons that influenced their decision to adopt IoT. Farmers reported IoT to be easily accessible while providing a relative advantage over previous systems. Farmers that had not adopted IoT expressed various reasons for doing so. For smaller farming operations, the complex nature of deriving sufficient benefits from IoT did not necessarily justify the costs. Other factors included the cultural change required to adopt new technology and external factors such as insufficient data coverage and governmental support.

 

Butale, M., Seymour, L.F. (2024). Understanding Limited Enterprise Systems Benefits in Government: An African Case. In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer View more...

Abstract
Enterprise systems (ES) are increasingly being used in government. Yet the benefits being realized are lower than anticipated and rarely match the level of investment. Post-implementation is when, through usage, benefits are realized and yet it is the ES phase researched the least. Hence there is little understanding of why benefits are limited in government departments. Therefore, this study aimed to answer the question ‘Why are limited ES benefits realized in government?’ The case studies chosen were in the government of an African developing country. Data included semi-structured interviews and secondary data. The explanatory model, inductively derived, reveals no ES benefits management processes in the cases. Low usage and limitations in their ES were predominantly due to the limited IT support post-implementation although limitations with the initial implementation project were raised. The root causes of limited benefits are organizational constraints within the government departments and poor consultant knowledge transfer.

 

Kumalo, M.O., Botha, R.A. (2024). POPIA Compliance in Digital Marketplaces: An IGOE Framework for Pattern Language Development. In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer View more...

Abstract
The rapidly growing digital marketplace landscape presents a unique challenge: reconciling the increased demand for consumer privacy with the innovative spirit and economic benefits these platforms offer. This tension is particularly acute in South Africa, where the Protection of Personal Information Act (POPIA) imposes stringent data privacy regulations. Striking a balance between these competing forces requires a nuanced approach that prioritises both consumer privacy and the continued growth of the digital marketplace ecosystem. This paper proposes a novel approach to navigating this complex landscape by leveraging the power of Pattern Languages. We utilize the IGOE framework (Inputs, Guides, Outputs, Enablers) to develop a POPIA compliance pattern language specifically tailored for digital marketplaces.

 

Magunje, C., Chigona, W. (2024). Educators’ Cybersecurity Vulnerabilities in Marginalised Schools in South Africa. In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer View more...

Abstract
Schools are experiencing an increased adoption and use of technology in curriculum delivery, administrative tasks, and community engagement. This was reinforced by the COVID-19 pandemic. The increased adoption has amplified exposure of schools to cyberattacks, as cyber criminals are finding opportunities such as deploying ransomware, stealing information, and extorting money from schools. Their positioning in a school setting makes educators paramount to cybersecurity in schools since cyber criminals can create a breach in the security system by targeting people, not computers. Schools in most Sub-Saharan Africa countries are, to an extent, defenceless to cyberattacks as the region faces various challenges such as limited financial resources, lack of or weak cybersecurity policies and comprehensive cyber safety initiatives, and limited cybersecurity knowledge and skills by educators. These challenges leave educators, the school and various stakeholders vulnerable to cyber-attacks which may compromise the security of the entire school information systems. This study aims to answer the question: What are the cybersecurity vulnerabilities of educators in marginalised schools in South Africa? The study employed a qualitative exploratory methodology using case studies of four schools located in marginalised schools in the Western Cape and Limpopo provinces of South Africa. We collected the data via semi structured interviews of educators. We analysed the data using thematic analysis based on the threat and coping appraisal constructs of the Protection Motivation Theory. The findings suggest that educators’ cybersecurity vulnerabilities emanate from a limited cybersecurity knowledge and skills which lead to compromised perception of threat severity, threat vulnerability, and misguided cybersecurity self-efficacy among educators. The study contributes to cybersecurity in education by emphasising the need for cybersecurity policies, interventions and initiatives for schools that can address the vulnerability and susceptibility of educators to cybersecurity threats.

 

Mbizo, T., Oosterwyk, G., Tsibolane, P., Kautondokwa, P. (2024). Cautious Optimism: The Influence of Generative AI Tools in Software Development Projects. In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer View more...

Abstract
Generative artificial intelligence has emerged as a disruptive technology with the potential to transform traditional software development practices and methodologies. This study examines the implications of integrating AI tools in software development projects, focusing on potential benefits, challenges, and perceptions of the broader software development community. The study employs a qualitative methodology that captures the sentiments and personal adaptive measures from a diverse group of industry professionals who integrate generative AI tools such as ChatGPT and GitHub’s Copilot in their software development projects. Findings suggest that generative AI tools aid developers in automating repetitive tasks, improve their workflow efficiency, reduce the coding learning curve, and complement traditional coding practices and project management techniques. However, generative AI tools also present ethical limitations, including privacy and security issues. The study also raises concerns regarding the long-term potential for job elimination (insecurity), over-reliance on generativeAI assistance by developers, generativeAI lack of contextual understanding, and technical skills erosion. While developers are optimistic about the positive benefits of generative AI use within project environments in the short term, they also hold a pessimistic view in the longer term. There is a need for the software development projects community to critically assess the use of generative AI in software development projects while exploring how to retain the critical aspect of human oversight and judgment in the software development process in the long term.

 

Mmango, N., Gundu, T. (2024). Cybersecurity as a Competitive Advantage for Entrepreneurs. In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer View more...

Abstract
This paper presents a systematic literature review focused on exploring the strategic role of cybersecurity as a competitive advantage for entrepreneurs. In the contemporary digital landscape, where cyber threats loom large, the ability to effectively manage and leverage cybersecurity practices has become a pivotal factor distinguishing successful entrepreneurial ventures. The review synthesizes existing research to identify actionable strategies through which cybersecurity can be transformed from a mere operational necessity into a significant competitive differentiator. Key findings from the literature underscore that robust cybersecurity measures enhance customer trust and loyalty, enable market differentiation, integrate seamlessly with strategic business objectives, foster innovation, and ensure compliance with regulatory standards. Strategies such as transparent communication about cybersecurity efforts, development of customized security solutions, integration of cybersecurity risk assessments into strategic planning, and investment in cybersecurity research and development are highlighted as effective means to leverage cybersecurity for competitive advantage. The review further elucidates the importance of adopting ethical data practices and staying abreast of regulatory compliance as mechanisms for reinforcing customer trust and navigating the complex legal landscape surrounding digital business operations. Through the analysis of selected case studies and best practices, the paper demonstrates practical applications of these strategies in real-world entrepreneurial contexts, illustrating how businesses can secure a competitive edge by prioritizing cybersecurity. Conclusively, the paper argues that cybersecurity, when strategically managed, offers entrepreneurs a unique opportunity to fortify their market position, enhance customer relationships, and drive sustainable business growth. It calls for a paradigm shift in how cybersecurity is perceived within the entrepreneurial ecosystem, advocating for its integration into the very fabric of business strategy development and execution.

 

Ntika, N., Chigona, W. (2024). Factors Affecting User Participation in the Design of Governmental Digital Services in South Africa.  In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer View more...

Abstract
User participation in the design of governmental digital services remain a latent policy stance in South Africa despite its benefits to address public service challenges in the country. Consequently, the country is marked by destructive service delivery protests due to government services that fail to meet the needs of the citizens. Citizens, as users of public services, may help improve the quality and quantity of public services by identifying problems with the existing digital services and suggesting new solutions that may better address their needs thereby improving user experience, satisfaction, and trust. This study, therefore, aimed to systematically identify and critically review relevant literature to provide insights into factors that affect user participation in the design of governmental digital service in South Africa. This study adopted a qualitative strategy and deductive approach and is underpinned by the revised IAP2‘spectrum of public participation’ as the theoretical framework. Sixteen articles, published from 2019–2024, from credible online databases, were included in this study. Findings show that factors that affect user participation in the design of governmental digital services include lack of trust in government, poor implementation of good governance, digital exclusion, to name but a few factors. Findings also show that the negative economic growth of the country has eroded government capacity to fully embrace inclusive governance. The study intended to broaden the understanding of factors that affect user participation in the design of governmental digital services in South Africa and to shape policies intended to address these factors.

 

Pieterse, H., Barbour, G., McDonald, A., Badenhorst, D., Gertenbach, W. (2024). Utilisation of a Virtual Honeynet to Proactively Secure the South African National Research and Education Network Against Cyberattacks. In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer View more...

Abstract
South Africa is witnessing a significant increase in cyberattacks. Although such an increase in cyberattacks can be attributed to various factors, poor investment in cybersecurity technology and lack of awareness are causing South Africa to be a target of interest. While cyberattacks are targeting various sectors, it is the cyberattacks impacting critical infrastructure that are a growing concern. The South African National Research and Education Network (SA NREN) is a high-speed network dedicated to science, research, education and innovation traffic. With the growth of the SA NREN and the continuous increase in cyberattacks affecting South African institutions, proactive steps are required to secure and protect the SA NREN. This responsibility lies with the SA NREN Cybersecurity Incident Response Team (CSIRT), which was established in 2016 to offer protection against cyberattacks. While various proactive measures are currently in place to monitor the SA NREN, the CSIRT continues to explore alternative cost-effective solutions to secure the NREN. This paper investigates the benefits of utilising a novel low-interaction secure shell (SSH) honeynet, referred to as the Virtual Honeynet, to monitor and proactively secure the SA NREN. The Virtual Honeynet uses virtual containers to reduce resource requirements and improve performance. The investigation involved the experimental deployment of theVirtual Honeynet on the SANRENover a twelve-day period and the evaluation of the captured data. The evaluation conducted focused on extracting behavioural and geographical intelligence from the raw data to guide the deployment of cyber measures to secure the SA NREN. The results presented in this paper confirm the value the Virtual Honeynet offers to the SA NREN as a technology to proactively secure the network.

 

Stamp, J., Mwapwele, S.D. (2024). Examining Data Governance to Determine How Democratic Data Management Can Be Achieved in Organizations. In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer View more...

Abstract
In this age of data-driven decision-making and a data-centric environment, robust data governance is vital for organisational performance and for upholding democratic and fair data practices. Managing data is increasingly challenging with large volumes across systems, evolving regulations, diverse stakeholder interests, and different implementation methods. The research gap is embedded in the realization that data governance is crucial in any organisation, but its implementation faces the demand need to address the horizontal relationships among data providers. The research question is how can democratic data governance (DDG) be implemented by institutions and organisations to better serve the interests of data providers? Employing a systematic literature review, this study is steered by the populations, exposures, and outcomes (PEO) framework to define the review’s boundaries. The databases used were Web of Science, ProQuest, and EbscoHost. The PRISMA framework was used to ensure a transparent, structured, and exhaustive examination of the selected literature. The inclusion criteria are articles in English, and peer-reviewed literature specifically about democratic data governance. Through thematic analysis, this study explores the obstacles and opportunities of DDG implementation in the 27 articles included. The key findings are grouped into three themes; data-driven societal challenges, aspects of data governance needed for democratic data governance, and challenges with adapting data governance. Addressing the horizontal relationship between data providers contributes to the policy as it engages with GDPR. The research contributes to the body of knowledge on data governance and the role of privacy.

 

Van der Merwe, T. M. (2024). An Exploratory Qualitative Study of Disruptive Technology Adoption Among Zimbabwean Small and Medium Enterprises in and Around Harare, Gweru, and Bulawayo. In: South African Computer Science and Information Systems Research Trends SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer View more...

Abstract
In taking advantage of the digital revolution, disruptive technology, which is capacitated by enhancements in computing capacity and Internet bandwidth, not only has the potential to overcome many challenges faced by small-to medium enterprises but also offers innovative opportunities to create new markets. Identified as a research gap, this paper outlines the state of disruptive technology adoption, estimates its perceived importance, identifies factors that affect its adoption, and gauges government support for providing a conducive environment in randomly selected rural and urban Zimbabwean SMEs. Semi-structured interviews with owners and staff of 12 rural and 12 urban SMEs in and around Harare, Gweru, and Bulawayo were conducted using a qualitative and exploratory approach. Results show its use to be limited, with only two rural and four urban SMEs using disruptive technology. Major factors which limit their use are poor mobile infrastructure, electricity, and the cost of services. Outside four, all urban respondents acknowledged the importance of disruptive technology, with three linking potential usefulness to their business processes. Rural respondents were less convinced, with responses ranging from uncertain, could be, to no need. Other factors identified were poor cash flow, limited knowledge, and no financial support. Despite policies in place, a total absence of government support was reported. The results paint a bleak picture of the current and potential trajectory of disruptive technology adoption in these Zimbabwean SMEs.

 

SAICSIT Online Proceedings Papers with Abstracts

Chibaya, C. (2024). Defining an Ontology for Ant-like Robots Based on Simulated Ant-agent Characteristics. In: SAICSIT 2024. Online Proceedings of the South African Institute of Computer Scientists and Information Technologists 2024 Conference.  View more...

Abstract
Swarm intelligence systems, where robotic devices are programmed with distinct abilities that operate at the individual level to produce collective emergent behaviour, are particularly promising in fields like nanotechnology. These systems are typically employed to solve complex real-world problems at minimal costs. For instance, ant colony systems emulate the behaviour of natural ants to tackle challenging issues. Solutions to complex optimization problems, such as the bridge crossing problem, vehicle routing problems, shortest path formation problem, and the travelling salesman problem, have been developed using this approach. This study draws inspiration from various simulated ant colony systems, exploring the low-level actions and capabilities of simulated ant-like robotic devices to develop an ant-bots swarm intelligence ontology. An ant-bot is conceived as a small, simple autonomous robot modelled after simulated ants. Individually, an ant-bot may not accomplish much, but as part of a swarm, these robots can generate impressive emergent behaviour. We examine the specific aspects of simulated ant agents that lead to emergent behaviour and incorporate these into the design of an ant-bots swarm intelligence ontology. Experimental tests identified three key components as the foundation of the desired swarm intelligence ontology. First, the swarm space component captures metadata about the configuration of the simulated environments, targets, and any global swarm rules. Second, the ant-bot. context emphasizes the individual abilities and activities of ant-bots. Lastly, the swarm interaction component details the communication mechanisms used, whether direct or indirect, local or global, inspired by nature, mathematical models, biological processes, or other methods. This swarm intelligence ontology serves as a formal knowledge representation model for ant-bots, encapsulating these aspects to enable effective swarm emergent behaviour

 

Grobler, G., Makura, S., Venter, H. (2024). A Technique for the Detection of PDF Tampering or Forgery. Online Proceedings of the South African Institute of Computer Scientists and Information Technologists 2024 Conference. View more...

Abstract
Wikipedia is a widely recognized and valuable source of information, However, it encounters persistent challenges in attracting and retaining active contributors. It is recorded that only 10 people from Zambia contribute and create content on wikipedia in the month of may 2023. While a large number consumes Wikipedia content, there is a noticeably low number of Wikipedians that contribute content on and about Zambia. This paper presents a Facebook plugin, WikiMotivate, aimed at motivating Zambian Wikipedians to update pre-existing content, add new entries, and share their natural expertise. WikiMotivate was implemented as a Facebook plugin that utilizes leaderboard and badge gamification features to encourage and incentivize active Wikipedia content contribution. Using a mixed-methods approach, historical Wikipedia edit histories were used to quantify content contributed by Zambian Wikipedians. In addition, user surveys were conducted to determine relative levels of awareness about Wikipedia, willigness to contribute to contribute content on Wikipedia and perceived motivating factors that affect content contribution on Wikipedia. Furthermore, a Facebook plugin, WikiMotivate, was implemented in order to be used an a service for motivating potential Zambian Wikipedians. Finally, in order to determine the most effective approach, a comparative analysis of leader-boards and badges was conducted with nine (9) expert evaluaters. The results clearly indicate that a significant proportion of Wikipedia content on and about Zambia is authored by Wikipedians from outside Zambia, with only 11% of the contributors, out of the 224, originating from Zambia. In addition, study participants were largely unaware of the various editing practices on Wikipedia; interestingly enough, most participants expressed their willingness to contribute content if trained. In terms of motivating factors, “Information Seeking and Educational Fullfilment” was as the key motivating factor. The Facebook plugin implemented suggests that incorporating leaderboards and badges is a more effective approach to motivating contributions to Wikipedia. This study provides useful insight into the landscape of Wikipedia content contribution in the Global South.

 

Herbst, C., Jeantet, L., Dufourq, E. (2024). Empirical Evaluation of Variational Autoencoders and Denoising Diffusion Models for Data Augmentation in Bioacoustics Classification. In: South African Computer Science and Information Systems Research Trends. SAICSIT 2024. Communications in Computer and Information Science, vol 2159. Springer. View more...

Abstract
Tampering or forgery of digital documents has become widespread, most commonly through altering images without any malicious intent such as enhancing the overall appearance of the image. However, there are occasions when tampering of digital documents can have negative consequences, such as financial fraud and reputational damage. Tampering can occur through altering a digital document’s text or editing an image’s pixels. Many techniques have been developed to detect whether changes have been made to a document. Most of these techniques rely on generating hashes or watermarking the document. These techniques however have limitations in that they cannot detect alterations to portable document format (PDF) signatures or other non-visual aspects, such as metadata. This paper presents a new technique that can be used to detect tampering within a PDF document by utilizing the PDF document’s file page objects. The technique employs a prototype that can detect changes to a PDF document, such as changes made to the text, images, or metadata of the said file.

 

Jideani, P., Gerber, A. (2024). Machine Learning-based NLP for Emotion Classification on a Cholera X Dataset. Online Proceedings of the South African Institute of Computer Scientists and Information Technologists 2024 Conference. View more...

Abstract
Recent social media posts on the cholera outbreak in Hammanskraal have highlighted the diverse range of emotions people experienced in response to such an event. The extent of people's opinions varies greatly depending on their level of knowledge and information about the disease. The documented research about Cholera lacks investigations into the classification of emotions. This study aims to analyze the emotions conveyed in social media posts regarding Cholera. A dataset of 23,000 posts was extracted and pre-processed. The VADER sentiment analyzer library was applied to determine the emotional significance of each text. Additionally, Machine Learning (ML) models were applied for emotion classification, including Long short-term memory (LSTM), Logistic regression, Decision trees, and the Bidirectional Encoder Representations from Tranformers (BERT) model. The results of this study demonstrated that LSTM achieved the highest accuracy of 76%. Emotion classification presents a promising tool for gaining a deeper understanding of the impact of Cholera on society. The findings of this study might contribute to the development of effective interventions in public health strategies.

 

Kavikairiua, J., Shava, F.B., Chitauro, M. (2024). Identifying Deep Learning Models for Detecting Child Online Threats to Inform Online Parents Education: A Systematic Literature. In: SAICSIT 2024. Online Proceedings of the South African Institute of Computer Scientists and Information Technologists 2024 Conference. View more...

Abstract
The COVID-19 pandemic has increased children's reliance on the Internet, exposing them to online risks, safety issues, mental health impacts, and educational inequities, and made it more difficult for parents to keep an eye on their children’s online activities. Most parents are unaware of online risks, despite advice against strangers, sharing sensitive information, and monitoring social circles, but often overlook the importance of ensuring a safe online experience. This paper presents deep learning models that can be used to educate parents on online child protection. In this paper, a systematic literature review was used to explore deep learning models for detecting child online threats to inform online parents. Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) were identified as appropriate deep learning models in this study and will in the future be used as a reference point to develop a deep learning model that can educate parents on protecting their children online. This study will have a significant impact on the field of cyber security since it is envisaged that the new ways of analyzing threats and proposing solutions to equip parents to protect children online proposed in this paper will enable parents to monitor their children's exposure and take necessary precautions in relation to trending threats. The model can also enable tracking of children's online interactions, identifying patterns, detecting risks, and understanding changes in online threats like cyberbullying. It is also superior for speech synthesis, question answering, and language modelling, among others, thus, parents can ask questions

 

Mazorodze, A.H. (2024). The integration of Software Engineering into IT Degree Programs: An analysis of the curriculum and industry relevance. In: SAICSIT 2024. Online Proceedings of the South African Institute of Computer Scientists and Information Technologists 2024 Conference. View more...

Abstract
Software Engineering is a core module in Information Technology (IT) degree programs. Because IT is ever evolving, the curriculum for Software Engineering should be updated more regularly to meet these technological changes. This research evaluates the Software Engineering curriculum at a selected private higher education institution in South Africa to fully understand the relevance of the content taught from a Software Engineer perspective. The study conducted a comprehensive review of the Software Engineering content taught at a selected private higher education institution in South Africa. Moreover, the study analyses the pedagogical approaches used to teach and assess Software Engineering students. Most importantly, the study solicits feedback through an online survey from the software engineering professionals in industry to ensure that graduates are well-equipped with the knowledge required to tackle problems in the real-world environment. A total of thirty-six (36) Software Engineers in South Africa took part in this study. Some of the current trends in Software Engineering include DevOps, Edge Computing, Microservices Architecture, Serverless Computing, Containerisation and Orchestration. The study established that technologies mentioned here should form part of the curriculum. The results of this study therefore offer insights for curriculum designers, educators, and policymakers to optimise the effectiveness of Software Engineering as a core module in Information Technology degrees. The study highly recommends that the curriculum be updated to incorporate DevOps, the Microservices Architecture as well as Serverless Computing among other contemporary technologies.

 

Mfati, E.S., Sigama, K., Langa, R.M., Moeti, N.M. (2024). An Accident-Avoidance Model for Driver Fatigue Detection using AIoT. n: SAICSIT 2024. Online Proceedings of the South African Institute of Computer Scientists and Information Technologists 2024 Conference. View more...

Abstract
Many of all traffic accidents are attributed to drivers who are less vigilant. This leads to a number of fatalities on our roads. While most drivers are aware of the risks associated with drinking and driving and texting while driving, many underestimate the hazards of driving when drowsy. There are very limited studied of fatigue detection using Artificial Intelligence of Things (AIoT).Systems for detecting driver weariness are being transformed by AIoT. According to literature, artificial intelligence systems use sophisticated algorithms and realtime data from IoT sensors to efficiently monitor driver behavior and spot signs of fatigue, lowering the likelihood of accidents on the road. The critical issue that a fatigue detection system must address is the question of how to detect fatigue accurately and early at the initial stage. This study explored the literature to find more on sensing and data collection, suitable machine learning algorithms for fatigue detection, and real-time monitoring to generate alerts.Finally the study developed a model and a flowchart for accident-avoidance

 

Oki, O.A., Agbeyangi, A.O., Mgidi, A. (2024). Blockchain in Healthcare: Implementing Hyperledger Fabric for Electronic Health Records at Frere Provincial Hospital.  In: SAICSIT 2024. Online Proceedings of the South African Institute of Computer Scientists and Information Technologists 2024 Conference. View more...

Abstract
As healthcare systems worldwide continue to grapple with the challenges of interoperability, data security, and accessibility, integrating emerging technologies becomes imperative. This paper investigates the implementation of blockchain technology, specifically Hyperledger Fabric, for Electronic Health Records (EHR) management at Frere Hospital in the Eastern Cape province of South Africa. The paper examines the benefits and challenges of integrating blockchain into healthcare information systems. Hyperledger Fabric’s modular architecture is harnessed to create a secure, transparent, and decentralized platform for storing, managing, and sharing EHRs among stakeholders. The study used a mixed-methods approach, integrating case studies and data collection methods through observation and informal questions, with the specific goal of understanding current record management methods and challenges. This method offers practical insights and validates the approach. The result demonstrates the role of blockchain in transforming healthcare, framed within a rigorous exploration and analysis. The findings of this study have broader implications for healthcare institutions seeking advanced solutions to address the persistent challenges in electronic health record management. Ultimately, the research underscores the transformative potential of blockchain technology in healthcare settings, fostering trust, security, and efficiency in the management of sensitive patient data

 

Rananga, S., Ngwenya, N., Mbooi, M., Isong, B., Matloga, P., Marivate, V. (2024). Misinformation Detection in Text for COVID-19 Healthcare Data in South Africa. In: SAICSIT 2024. Online Proceedings of the South African Institute of Computer Scientists and Information Technologists 2024 Conference. View more...

Abstract
The COVID-19 pandemic has witnessed an alarming expansion of misinformation, posing critical threats to public health. This research focuses on detecting misinformation within text data sourced from South African healthcare datasets during the COVID-19 crisis. The study aims to improve automated misinformation detection models tailored to the South African context by leveraging natural language processing (NLP) techniques and machine learning algorithms such as Logistic regression (LR). The investigation involves extensive analysis of datasets containing COVID-19 healthcare-related text, including the training and evaluation of machine learning models. In addition, NLP techniques, including LR, will be utilized to extract key features indicative of misinformation. The findings are poised to advance the development of effective tools and strategies to combat misinformation. Moreover, the study will adopt an inclusive approach by conducting analyses in both English and low-resource languages like isiZulu. This is critical to enhancing public health communication and strengthening defences against the potential resurgence of COVID-19, thereby safeguarding public well-being

 

Tait, R., Vogts, D., Greyling, J. (2024). A Technology Based Model for Problem-Solving Skills Development in the Intermediate Phase. In: SAICSIT 2024. Online Proceedings of the South African Institute of Computer Scientists and Information Technologists 2024 Conference. View more...

Abstract
An educational model to improve problem-solving skills among learners is introduced. The literature review emphasises the importance of these skills, the impact of smart classrooms, and the influence of colour on learning. The model offers guidance for educators and learners to foster skills development in the intermediate phase through a structured workflow. A set of metrics to aid in the selection of tools to improve problem-solving skills is proposed. Based on the model, a proof-of-concept Internet of Things based system was successfully implemented and evaluated using a metric-based evaluation based on the proposed metrics. This evaluation highlights the effectiveness of the model in the implementation of a system in limited infrastructure and diverse educational settings.

 

Worthington, C., Holtzhausen, L., Kuttel, M. (2024). User-centred design and development of a web-based Western Cape substance use assessment tool (WC-SUDAT).  In: SAICSIT 2024. Online Proceedings of the South African Institute of Computer Scientists and Information Technologists 2024 Conference. View more...

Abstract
Substance use disorders (SUDs), the uncontrolled use of substances despite harmful consequences, is a significant problem in South Africa, especially in the Western Cape. An important component in the fight against SUDs are questionnaires to assess the risk of an SUD that are administered by social workers to identify targeted interventions. A web-based questionnaire with automated aggregation of responses can reduce the administrative burden placed on social workers. Here we use a user-centred design approach to build a web-based substance use disorder assessment tool localised to the Western Cape: WC-SUDAT. Our three-phase User Centred Design methodology comprised a first prototype; followed by evaluation of its suitability through a contextual inquiry, a usability test and heuristic evaluation; and then implementation of a final prototype incorporating unanticipated features critical for field use that were identified in the evaluation. This process was effective in generating a final prototype webtool with a dual function as both an SUD assessment tool and an organisational management tool. This deployment-ready prototype is a better fit for the needs of NGOs working with substance abuse disorders than our original conception of the webtool, thus validating a User-Centred design approach

 

Birtles, K., Davids, Z., Seymour, L. (2024). Constraints and Outcomes of Giving Free Wi-Fi Access to Learners: A South African Case Study..  In: SAICSIT 2024. Online Proceedings of the South African Institute of Computer Scientists and Information Technologists 2024 Conference. View more...

Abstract
We cannot imagine a world without the Internet both within our personal life and for educational purposes. In South Africa, the government’s “Smart Classroom project” introduced Wi-Fi in schools. However, its adoption by learners has been slow which has been attributed to school management concerns. This research aimed to describe the constraints and perceived outcomes of giving free Wi-Fi access to learners to understand the slow adoption. This qualitative multiple case study in five Western Cape schools used a hybrid inductive-deductive analysis of interviews and documents. The study identified nine constraints preventing learners accessing the Wi- Fi, dominated by school restrictive policies and lack of device control. The study also highlighted both positive and negative outcomes for learners using the school’s Wi-Fi. Game-based learning in classrooms and accessing online educational content are some positive outcomes. Negative outcomes are being distracted from learning, cyberbullying, accessing pornography and internet abuse. This study will be of particular interest to education organizations and government decision makers highlighting areas of concerns amongst school management when providing free Wi-Fi access to learners. Finally, the paper suggests that the Western Cape Education Department (WCED) considers increasing their financial investments into procuring additional bandwidth, additional teacher training, and smart applications to help manage personal devices of learners and staff.

 

Bunt, L., Taylor, E., Greeff, J. (2024). Stakeholder Theory and Enterprise Architecture in Serious Games: An Integrative Review.  In: SAICSIT 2024. Online Proceedings of the South African Institute of Computer Scientists and Information Technologists 2024 Conference. View more...

Abstract
This study addresses the oversight of stakeholder-focused strategies in traditional serious game frameworks by proposing a stakeholder-centred approach to serious game design. Through a comprehensive review and synthesis of literature across serious games, stakeholder theory, and enterprise architecture, this paper highlights the importance of integrating stakeholder theory and enterprise architecture principles in the development of serious games. An integrative literature review process—screening, extracting, synthesizing, and evaluating data—led to the development of a conceptual framework for serious game design. This framework is characterized by its flexibility, stakeholder-centeredness, goal orientation, and supportive nature, designed to overcome common challenges in serious game development. It prioritizes the needs and interests of all stakeholders, including players, designers, investors, and regulators, emphasizing the application of stakeholder theory and enterprise architecture methods to enhance serious game development. The proposed framework facilitates the prioritization of stakeholder needs and alignment with organizational goals, thereby improving player experiences and ensuring scalability and security

 

Chikotie, T.T., Watson, B., Watson, L.R., Banda, T.V. (2024). Strategies for Healthcare Resilience: A Comparative Evaluation of ARIMA and LSTM Models in Predicting COVID-19 Hospital Admissions. Online Proceedings of the South African Institute of Computer Scientists and Information Technologists 2024 Conference. View more...

Abstract
The Eastern Cape Province faced significant challenges in hospital bed planning due to high COVID-19 infection rates and a lack of effective estimation models to support decision-making. This was because the first wave was characterised by the ancestral strain with a mutation at position Asp614Gly, the second by the beta variant (B.1.351), the third by the delta variant (B.1.617.2), and the fourth by the omicron variant (B.1.1.529). Therefore, there is a need to develop adequate short-term prediction models for forecasting the number of hospital admissions. This study proposed a comparative analysis of univariate (Autoregressive Integrated Moving Average) ARIMA, (Long-term Short Memory) LSTM, and ensemble ARIMA-LSTM models for COVID-19 hospital admissions forecasting. Leveraging historical data, we evaluated model performance in the public and private sectors using (Root Squared Mean Error) RSME, (Mean Absolute Error) MAE, and R-squared error. Our findings revealed that the LSTM model is the better performing model for both the public - RMSE: 0.146963, MAE: 0.1018066, R2: 0.9990176 and private sectors - RMSE: 4.2412125, MAE: 0.0816642, R2: 0.9990707. LSTM has the lowest RMSE and MAE values, indicating more accurate predictions, and the highest R² values, indicating the best fit to the actual data. The ensemble ARIMA-LSTM improves performance over the ARIMA model but is still not as good as the LSTM model. The ARIMA model has the poorest performance among the three models for both sectors. These models were also compared with other existing models from previous related studies. Due to its proven resilience and heightened predictive precision, using LSTM holds promise for enhancing pandemic forecasting, thereby facilitating improved hospital admissions planning and management strategies.

 

Kaisara, G., Yakobi, K., Peel, C., Mare, A. (2024). “System e Down”: Citizens’ Perceptions of the Failures of e-Government Systems in Botswana. In: SAICSIT 2024. Online Proceedings of the South African Institute of Computer Scientists and Information Technologists 2024 Conference. View more...

Abstract
Globally, the importance of e-governance platforms in delivering services to citizens has significantly changed the way in which elected officials, public servants and ordinary people interact. However, in the African context, the digitisation of public services raised numerous unfulfilled expectations of seamless electronic service delivery among the public which often has to endure long queues, filling of countless forms and travelling long distances. Given that digitisation, datafication and platformisation processes are predominantly driven by platform companies from the Global North, most e-government platforms and systems are not easily relatable to the African context. Because of the inbuilt colonial matrix of power, African governments are too dependent on hardware and software from Asia and North Africa. This article relies on virtual ethnography and qualitative context analysis to assess user perceptions of e-government platforms and system failures in Botswana. Users’ perceptions and sentiments of e-government system failures were extracted from both government and corporate social media Facebook pages announcing e-government system failures in Botswana. Thematic analysis was adopted, resulting in the emergence of four themes, namely; encouraging law-breaking, loss of confidence in government, exasperation and security concerns. The results are discussed and implications put forward.

 

Kroeze, J.H. (2024). Pointers for Ubuntu Information Systems Ethics. In: SAICSIT 2024. Online Proceedings of the South African Institute of Computer Scientists and Information Technologists 2024 Conference. View more...

Abstract
Although there has been interest in the ethical aspects of information systems (IS) since the 1980s, various authors have recently lamented the fact that not enough research has been done in the area and that ethics is often ignored in the IS industry. When one searches for research on the decolonisation and Africanisation of IS ethics, few outputs can be found. The main research question that this article addresses is: How can African knowledge systems and ways of knowing inform and enrich IS ethics? The main aim of the article is to identify appropriate ethical insights, borrowed from Ubuntu-informed ethics, information ethics and business ethics to serve as pointers for Ubuntu IS ethics. The study is a conceptual study which follows a philosophical approach. The research is a rudimentary attempt to enrich IS ethical theory from an African viewpoint. The most important contribution of the paper is the proposal of a root for an Ubuntu-based IS ethic which could be used to counteract the hegemony of Eurocentric values embedded in information and communication technology

 

Mahlasela, O., Baloyi, E., Siphambili, N., Khan, Z.C. (2024). Artificial Intelligence Impact on the realism and prevalence of deepfakes. In: SAICSIT 2024. Online Proceedings of the South African Institute of Computer Scientists and Information Technologists 2024 Conference. View more...

Abstract
Deepfakes, synthetic media manipulated by Artificial Intelligence (AI), have become a growing concern in the information landscape. This paper explored the impact of AI on the realism and prevalence of deepfakes. Therefore, this study examined how AI advancements in machine learning and generative models have facilitated the creation of increasingly convincing deepfakes. The analysis looked at the rise of hyper-realistic deception and the societal impact of deepfakes. In addition to recognizing the challenges, a framework was developed for the detection of deepfakes. Finally, this study discussed the potential mitigation strategies, such as the development of deepfake detection tools and fostering media literacy.

 

Malanga, D.F., Chigona, W. (2024). A Model for Predicting Continuance Intentions of mHealth with Community Health Workers in Malawi: From User Expectations Perspective.  In: SAICSIT 2024. Online Proceedings of the South African Institute of Computer Scientists and Information Technologists 2024 Conference. View more...

Abstract
This paper aims to model user expectations as predictors of continuance intentions of mHealth with community health workers (CHWs) in Malawi, a developing country context. The study extends the expectation confirmation model to include effort expectancy, and quality triads (system quality, information quality, and service quality). A survey questionnaire was used to collect data from 176 randomly sampled CHWs in three district health facilities in Malawi. Partial least squares method to structural equation modelling (PLS-SEM) was used to analyse data. The study found that effort expectancy, confirmation, satisfaction, and post-usage usefulness had a significant influence on CHWs’ continuance usage intentions with mHealth in Malawi. However, the unexpected results were that quality triads did not yield positive effects on the continued usage intentions of CHWs with mHealth. This was contrary to the established IS extant literature. The study concludes by making recommendations for policy and research practice.

 

Waiganjo, I.H., Osakwe, J., Azeta, A. (2024). The Four Processes of an Effective Cyber Security Policy. In: SAICSIT 2024. Online Proceedings of the South African Institute of Computer Scientists and Information Technologists 2024 Conference. View more...

Abstract
This article explores the process of cybersecurity policy formulation, implementation, and modification, emphasising the critical role of policy compliance in fortifying organizational digital defences. Drawing insights from various literature sources, the article highlights the multifaceted nature of cybersecurity policies, encompassing technological, procedural, and humancentric elements. The policymaking steps, including formulation, implementation, modification, and compliance, are described, underscoring the importance of tailoring policies to unique organizational cyber platforms. The study identifies and elaborates on essential cybersecurity policies, such as privacy, email security, net- work security, Wi-Fi usage, physical security, password management, and incident response. The article also introduces Lubua and Pretorius's cyber-security policy framework, illustrating seven key entries for comprehensive policy development. Furthermore, it stresses the ongoing need for policy compliance as a cornerstone for effective cybersecurity within organizations, involving both technical and non-technical solutions. The dynamic nature of technology and the continuous evolution of cyber threats necessitate periodic reviews and modifications to cybersecurity policies. The iterative process of cybersecurity policy development, implementation, compliance, and modification establish a robust frame- work, safeguarding digital infrastructure and enabling effective responses to evolving cyber challenges.

 

Young, J., Pekane, A., Kautondokwa, P. (2024). Behavioural Predictors that Influence Digital Legacy Management Intentions among Individuals in South Africa. In: SAICSIT 2024. Online Proceedings of the South African Institute of Computer Scientists and Information Technologists 2024 Conference. View more...

Abstract
An emerging phenomenon, digital legacy management explores the management of digital data individuals accumulate throughout their lifetime. With the integration of digital systems and data into people's daily lives, it becomes crucial to understand the intricacies of managing data to eventually become one’s digital legacy. This can be understood by investigating the significance of behavioural predictors in shaping digital legacy management. The objective of this study is to explore how behavioural predictors influence the intentions of individuals in South Africa towards managing their digital legacy. This entailed: 1) investigating the impact of attitude, subjective norms, and perceived behavioural control on these intentions; 2) exploring the perceived usefulness of digital legacy management systems; and lastly 3) understanding the implications of response cost and task-technology fit on individuals' inclinations towards digital legacy planning. Data were collected (n = 203 valid responses) from South African residents using an online survey and analysed using partial least squares structural equation analysis (PLS-SEM). Results indicate that attitudes, peer opinions, personal resources and skills are significant positive influences on digital legacy management intention. Recognizing and understanding these behavioural predictors is key when developing region-specific and culturally sensitive digital legacy management tools, awareness campaigns and policies. Furthermore, it could pave the way for more tailored strategies, ensuring effective transfer of post-mortem data, reducing potential conflicts, and providing clarity when dealing with post-mortem data

 

 




© 2024 Nelson Mandela University    WCMS