Research Articles (Informatics)

Permanent URI for this collectionhttp://hdl.handle.net/2263/1697

Browse

Now showing 1 - 20 of 295

The use of heat maps to explore perceived visual indicators of online fake news
(Elsevier, 2024) Matthee, Machdel C.; Hattingh, Maria J. (Marie); Normund-van Wyk, Ludwig; machdel.mathee@up.ac.za
The study uses heat maps to explore visual indicators of online fake news as perceived by news consumers. Participants in this study were asked to browse through three separate documents, two of which contained fake news, which they were not made aware of. After browsing the three documents, participants were asked a series of questions via a questionnaire relating to the three articles’ content. These answers, in combination with the participants’ combined heat map data, were analysed to determine what participants consider as visual indicators of online fake news. Participants attempted to establish the source’s credibility and evaluate the indicators provided in the articles. It appeared that the title, URL and website names were important indicators. In addition, images attracted attention, too many adverts were considered a red flag, and an organized layout with an author, date, sources, and secure website increases the credibility of the source.
Student intentions to continue with distance learning post-COVID : an empirical analysis
(Public Library of Science, 2024-01-17) Steyn, Adriana Aletta (Riana); Van Slyke, Craig; Dick, Geoffrey; Twinomurinzi, Hossana; Amusa, Lateef Babatunde; riana.steyn@up.ac.za
The aftermath of COVID changed how students learn, mainly moving to a distance learning model. The research reported in this paper investigated the organizational and individual factors that influence the preference for continuing with distance / online learning post-COVID. Partial least squares structural equation modeling was applied to a model developed for this research, based on data from 452 students from residential universities in South Africa. The key results reveal an overall reluctance to continue with distance learning. This is despite the technological and faculty support offered to university students and how distance learning fits their learning styles. This is likely due to student living conditions and their perception of low institutional concern. On the other hand, faculty support has a more substantial impact on continuance, compared with a generally negative perception of support from the universities. The research underscores the importance of addressing student reluctance to continue with distance learning by improving institutional support and tailoring learning styles. The research enhances our understanding of crucial factors influencing students’ preference for distance/online learning post-COVID. It also underscores the pedagogical shifts brought about by the pandemic, particularly highlighting the changing roles of faculty support and the impact on students’ living conditions.
A bibliometric analysis review : the emerging technology of artificial intelligence for non-bio inspired and bio-inspired algorithm of wireless sensor network from 2005–2022
(Cerebration Science Publishing, 2025-02) Aroba, Oluwasegun Julius; Rudolph, Michael; Naicker, Nalindren; Karodia, Khadija; Gupthar, Avintha; Bugwandin, Vinay; Ramchander, Manduth; Adeliyi, Timothy
Rapid developments in technology, business, and social norms have been observed in the twenty-first century. The fourth industrial revolution has been brought about by most industries moving toward automation and reducing human intervention. Wireless sensor networks are incredibly important to the fourth industrial revolution since they help with modernization. WSNs are networks of sensor and routing nodes that can be integrated into a variety of control systems, such as those used for home automation, electric-power automation, and environmental monitoring. A key problem that typically afflicts wireless sensor networks is node localization (WSNs). As a result, several algorithms, to ameliorate the challenges WSNs confront, both bio-inspired and non-bio-inspired solutions have been presented. From 2005 through 2022, the Scopus database was searched for publications. WSNs are used in published research paper statistical analysis, Microsoft Excel 365, VOSviewer, RStudio, and Biblioshiny packages were used. For this seventeen-year study period, a total of 36,377 published documents were in the Scopus database. 765 papers in all were examined following the implementation of the exclusion criteria. This study highlights the global research production of bio-inspired and non-bioinspired algorithms in wireless sensor networks, together with their status and tendencies. It can assist IoT and wireless sensor network researchers in gaining a thorough understanding of the most advanced algorithms in this area.
Investigating the Internet skills of older adults in South Africa
(CHRIST University, 2024-10-10) Gwala, Nkonzwenhle; Mawela, Tendani; tendani.mawela@up.ac.za
The older adult population is growing in South Africa. However, it is reported that approximately only 3.6% of older adults use the Internet. Using information technology resources and the internet can enable older adults to be independent for longer, and the internet may improve older adults’ quality of life. Additionally, access to technology and the internet supports access to information and digital inclusion. This study investigated older adults’ level of internet skills in South Africa and their experiences regarding internet use. The research followed a positivist, quantitative approach and used online questionnaires for data collection. The results showed that older adults appear to be lacking in internet skills in South Africa. The overall mean scores were low, where mobile skills seemed to be the lowest-rated skill for the respondents. The evaluation also showed that online operational and creativity skills appeared low for older adults.
Navigating cultural tensions and adaptation across changing global IT work environments : a case study of a South African IT unit reorganization
(Routledge, 2025) Naidoo, Rennie; Van der Merwe, Jean-Pierre
Navigating cultural tensions during a reorganization to a globally distributed IT unit is a formidable challenge for IT managers and workers across diverse cultural contexts. Our case study integrates and extends national cultural dimensions and organizational culture models in global IS research with the relational dialectics perspective to explore the discursive and material nature of cultural tensions experienced by local IT unit workers in a multinational technology organization. Drawing from in-depth interviews, we use thematic and contrapuntal analysis to examine the cultural tensions onsite local IT unit managers and specialists from South Africa expressed about their changing relations with an offshore global IT headquarters. Our results reveal an overarching autonomy versus control tension and three interrelated subtensions—workplace flexibility versus workplace standardization, strategic position versus compliant position, and empowerment advocacy versus bureaucratic dictates—which influenced contradictory meanings and responses to the reorganization. These tensions and adaptation strategies, though rooted in the South African IT unit context, resonate with global IT work environments, particularly in emerging economies and high- and low-power-distance cultures like China and Western Europe, where similar dynamics of cultural negotiation between local autonomy and global control are evident. The central contribution of this research is a global IT cultural tensions model, offering practical implications for fostering more inclusive and harmonious IT work environments across diverse cultural and regional contexts.
Systematic literature review on opportunities for digital innovation in the healthcare sector
(Koegni-eHealth Innovation for Development, 2024-03) Osembe, Lufungula; Adebesin, Funmi; Smuts, Hanlie
BACKGROUND AND PURPOSE: The concept of digital innovation has garnered significant attention recently and has the potential to revolutionise industries. With the recent advancements in healthcare and the emergence of new digital technologies, this manuscript reports the results of a systematic literature review (SLR) aimed at identifying the opportunities that digital innovation can afford the healthcare sector. These opportunities could guide policymakers and healthcare practitioners in making informed decisions regarding the implementation of digital innovation in healthcare. METHODS: This study conducted a systematic literature review (SLR) that involved defining the purpose of the literature review, establishing inclusion and exclusion criteria, identifying data sources and search strategies, screening and determining eligibility, and making inclusion/exclusion decisions. This study also included quality assessment criteria to evaluate the selected publications. 1478 research papers were analysed using descriptive statistics and thematic analysis to identify themes from the papers included in the SLR. RESULTS: The findings presented nine themes related to opportunities for digital innovation in healthcare. The following themes were discussed: precision/personalised medicine/healthcare, interoperability, privacy, and security, improved decision-making process, real-time/remote monitoring, enabling new processes/services/products, training/education, knowledge acquisition/sharing, and other themes that could not be categorised. These themes include cost, effectiveness, and time management. CONCLUSIONS: This study utilised thematic analysis to examine the themes that emerged from the findings. These themes centred around the opportunities for digital innovation in healthcare. Awareness about each opportunity for digital innovation in healthcare is discussed. This study discussed the limitations of this SLR and future research opportunities for policymakers and healthcare practitioners wanting to implement digital innovation in healthcare.
A reflective thematic analysis into the perceptions of pregnant radiographers regarding the usefulness of the Pregidose mobile app to enhance fetal dosimetry and well-being : qualitative exploration
(JMIR Publications, 2024) Essop, Hafsa; Kekana, R.M. (Mable); Brosens, Jacques; Smuts, Hanlie
BACKGROUND : Pregnancy apps are widely used by pregnant women, who benefit from self-tracking features to support their health goals. Pregnant radiographers are considered a high-risk group of health workers practicing in ionizing radiation environments. Radiation exposure above threshold limits can cause harmful genetic effects on a fetus. Accordingly, pregnant radiographers are required to wear special fetal dosimeters, which provide real-time readings of radiation dose exposure to the fetus. Pregnant radiographers have the responsibility to self-track their fetal doses to ensure that the threshold limit of 1 mGy is not exceeded. The traditional method used to track doses includes a written log of doses in a notebook. Thus, PregiDose, a unique offering in the context of pregnancy apps, was developed to enhance fetal dose tracking and monitoring using technological methods. OBJECTIVE : This study aims to describe the users’ perceptions of the app’s usefulness using PregiDose in a natural setting. METHODS : The overarching framework adopted for the study was a design science research (DSR) methodology encompassing five steps, namely (1) problem awareness, (2) suggestion, (3) development, (4) evaluation, and (5) conclusion. This paper presents the evaluation step of DSR. DSR step 4 included a qualitative approach to explore users’ perceptions regarding the app. Data were collected using a semistructured interview guide. Open-ended questions were guided by the app’s core features, namely dose tracking, education, and wellness. In total, 17 pregnant radiographers in South Africa enrolled to use the app, 9 (53%) engaged with the app, and 4 (24%) agreed to participate in the feedback interviews. The data were collected from October 2023 to March 2024 and analyzed using a reflective thematic data analysis method. RESULTS : Three overarching themes emerged from the data, namely (1) usefulness of PregiDose, (2) barriers to PregiDose adoption and use, and (3) recommendations for the advancement of PregiDose. Users labeled the app’s usefulness as positive and perceived it as a modern approach to traditional dose-tracking methods. They perceived the graph output of the dose-tracking feature to be useful for viewing their accumulative doses. They did not fully engage with the journaling feature, indicating that it was a personal preference and not a practice they would usually engage in. Physiological barriers, such as fatigue and “pregnancy brain,” were contributors to decreased engagement. Finally, because of the demanding workload and fast-paced nature of the radiography department, users recommended the automation of fetal dosimetry through the Internet of Things. CONCLUSIONS : PregiDose is an occupational health and safety mobile app developed for pregnant radiographers through a DSR approach. The app offers a modern method of dose tracking consistent with technological advancements in the context of self-tracking. However, future implementation would require using Internet of Things to make fetal dose tracking more effective.
Embedding sustainability : sociotechnical knowledge management guidelines for digital decarbonization in the Society 5.0 era
(MDPI, 2025-02) Smuts, Hanlie; Van der Merwe, Alta; hanlie.smuts@up.ac.za
Economic, social, and environmental sustainability emphasizes the need for organizations to integrate sustainability strategies into their core business and business development plans. The era of Society 5.0 is characterized by human-centeredness and digital leadership. It requires embedding sustainability practices and Green Information Technology (IT) while leveraging human–technology relationships to promote social good. However, embedding these practices into organizational culture is challenging due to resistance to change and the need for widespread mindset shifts. This study selected a focus group of eight South African participants to define sociotechnical knowledge management (KM) guidelines for embedding sustainable practices in organizations to promote digital decarbonization aligned with the Society 5.0 vision. Our findings suggest ten elements for the guidelines to incorporate: external environment, organizational context, business drivers, business outcomes, monitoring and evaluation, KM processes, technology enablers, sociotechnical KM tactics, knowledge assets, and execution considerations. By adopting such guidelines as a sustainability strategy, organizations can integrate KM practices into the human-centered and cyber-physical philosophy of Society 5.0. This approach aligns employee behavior with technological tools, enabling organizations to make data-driven decisions, reduce digital waste, and foster a culture of environmental responsibility. In addition, this approach enhances collaboration and innovation, benefiting all stakeholders and advancing sustainable development.
A transformer-based approach to Nigerian Pidgin text generation
(Springer, 2024-12) Garba, Kabir; Kolajo, Taiwo; Agbogun, Joshua B.; taiwo.kolajo@up.ac.za
This paper describes the development of a transformer-based text generation model for Nigerian Pidgin also known as Naijá, a popular language in West Africa. Despite its wide use, Nigerian Pidgin remains under-resourced, particularly in areas related to text generation and natural language processing. These difficulties are primarily due to technological constraints rather than the language’s fundamental attributes. There is currently a demand for Nigerian Pidgin-specific solutions because it is used in everyday communication and has a unique linguistic blend. This paper aims to close this gap by exploring the application of state-of-the-art transformer technology to develop a text generation model for Nigerian Pidgin. This work uses the public Afriberta-corpus dataset to optimize the Generative Pre-trained Transformer (GPT-2) model across a sizeable dataset. The performance evaluators, BLEU and Perplexity metrics provide a detailed breakdown of the model’s text quality and predictive accuracy. Despite the difficulties caused by a limited amount of training data, preliminary evaluations show that the model can generate coherent Nigerian Pidgin text. The performance evaluation yielded perplexity scores of 43.56 for variable target reference length and 43.26 for fixed text length. BLEU scores of 0.15 for fixed max length and 0.56 for variable reference target length. This highlights the quality of generated text and the significant improvement when the generated text length is aligned with the reference target. Our work was benchmarked against African American Vernacular (AAVE) revealing that BLEU scores for AAVE are significantly lower than those for Standard American English, with BLEU given as 0.26. Our Nigerian Pidgin model, with a BLEU score of 0.56, shows a better performance. However, both results suggest that both dialects are challenging for language models. Leveraging the pre-trained transformer-based language model and evaluation metrics, we showcase the model’s capacity for coherent Nigerian Pidgin text generation. For future research, the research work can serve as a good foundation for advancement and progress in the Nigerian Pidgin language generation and other low-resource languages.
Open source software adoption in the financial services industry : exploratory evidence from South Africa
(IGI Global, 2024-01) Langerman, Josef; Daramola, Olawande
Although OSS has gained traction across industries worldwide, few studies have been reported on OSS adoption in the financial industry and even less in the context of developing countries. This paper presents an exploratory overview of the state of OSS adoption in the financial industry using the case of South Africa as a country with the largest and most developed financial sector in Africa. To achieve this, we conducted a qualitative study that leveraged the Technology-Organisation-Environment (TOE) framework, and the Diffusion of Innovation (DOI) theory as theoretical references and used thematic analysis to analyse the data collected from a focus group discussion (FGD) of eight experts from four financial services organisations based in South Africa. The study's findings reveal the state of practice of OSS and the technological, organisational, and environmental factors that affect OSS adoption, and diffusion in financial services organisations in South Africa. We also identified seven research themes that should gain the attention of researchers from now on.
Harnessing disruptive innovation : a conceptual model for SME growth and adaptation
(MDPI, 2024-11-27) Smuts, Hanlie; Weilbach, Elizabeth Helena (Lizette); Padachi, Kesseven; Boolaky, Aleesha; Mauree-Narrainen, Diroubinee; Ramphul, Needesh; Chittoo, Hemant
In the context of sustaining innovation, small and medium enterprises (SMEs) strive to enhance their market position through product improvements. However, globalization and rapid technological advancements pose significant challenges, urging SMEs to integrate innovative capabilities into their business models. Effective SME business model innovation, fueled by customer insights, process efficiency, and technology application, can promote development and unlock additional value-creation resources. Despite this, SMEs encounter difficulties in capitalizing on disruptive innovation due to the competitive, technology-driven, and volatile global market. This study aimed to define a comprehensive conceptual model of disruptive innovation specifically tailored for SMEs. Through an automated content analysis of relevant literature, 13 themes and 82 concepts were identified and categorized into four strategic alignment domains. These domains provide a framework for operationalizing the findings and constructing a conceptual model of disruptive innovation. Utilizing this conceptual model as a checklist can assist SMEs in turning disruption into opportunity, thereby supporting their adaptation and growth in an increasingly competitive landscape.
Applications, technologies, and evaluation methods in smart aquaponics : a systematic literature review
(Springer, 2025-01) Anila, Mundackal; Daramola, Olawande; wande.daramola@up.ac.za
Smart aquaponics systems are gaining popularity as they contribute immensely to sustainable food production. These systems enhance traditional farming with advanced technologies like the Internet of Things (IoT), solar energy, and Artificial Intelligence (AI) for increased proficiency and productivity. However, assessing the performance and effectiveness of these systems is challenging. A systematic literature review (SLR) was conducted to examine the applications, technologies, and evaluation methods used in smart aquaponics. The study sourced peer-reviewed publications from IEEE Xplore, Scopus, SpringerLink and Science Direct. After applying inclusion and exclusion criteria, a total of 105 primary studies were selected for the SLR. The findings show that aquaponics predictions (27%) have been under-explored compared to applications that involved monitoring or monitoring and controlling aquaponics (73%). IoT technologies have been used to create prototype aquaponic systems and collect data, while machine learning/deep learning (predictive analytics) are used for prediction, abnormality detection, and intelligent decision-making. So far, predictive analytics solutions for aquaponics yield prediction, return-on-investment (ROI) estimates, resource optimisation, product marketing, security of aquaponics systems, and sustainability assessment have received very little attention. Also, few studies (37.7%) incorporated any form of evaluation of the proposed solutions, while expert feedback and usability evaluation, which involved stakeholders and end-users of aquaponics solutions, have been rarely used for their assessment. In addition, existing smart aquaponics studies have limitations in terms of their short-term focus (monitoring and controlling of aquaponics not undertaken over a long time to assess performance and sustainability), being conducted mostly in controlled settings (which limits applicability to diverse conditions), and being focused on specific geographical contexts(which limits their generalizability). These limitations provide opportunities for future research. Generally, this study provides new insights and expands discussion on the topic of smart aquaponics.
Capturing memories : an engaged research project using community digital storytelling
(Common Ground Research Networks, 2023-06) Jordaan, Martina; Mennega, Nita; martina.jordaan@up.ac.za
This article reports on a community digital storytelling project and explores how a digital storytelling project with a local community creates an accessible repository of information about the previously unrecorded history of the community. The storytelling project encourages the community to tell their stories or discuss issues that are important to them. The program has been running for more than twelve months, with regular training in cloud-based video editing software. The storytelling videos are uploaded on the community organization’s YouTube channel. After each training session, the participants complete a survey reflecting on their experience with creating the videos. Findings indicate that accessing technology is crucial for personal empowerment and community development. It allows individuals to express themselves, challenge dominant narratives, and shape their own destinies. By creating digital stories, individuals can gain self-confidence and belief in their potential. This belief is a powerful tool for personal and community growth, enabling individuals to take risks, pursue their goals, and make meaningful changes in their lives and society.
Sensory gardens in under-resourced early childhood development centers
(Common Ground Research Networks, 2025-06) Jordaan, Martina; Falk, Audrey F.
The objective of this study is to investigate the possible incorporation of a sensory garden in early childhood development (ECD) centers located in Mamelodi, a township adjacent to Pretoria, South Africa. This article reports on efforts to empower ECD practitioners to provide children with opportunities to interact with nature. The study involved a five-week course with brief weekly sessions for thirty ECD center practitioners. The participants were provided resources to develop a small sensory garden at their ECD centers. The researchers visited fourteen ECD centers whose staff indicated they had created a sensory corner. The visits were used to determine if the ECD center practitioners applied the acquired knowledge to develop a small sensory garden that could be used for teaching and learning. Thereafter, the participants completed a reflective survey on the course, the process, and the challenges they experienced. The study revealed that additional basic training was required to encourage ECD practitioners to integrate sensory elements into preschools. While many ECD centers were able to introduce fundamental sensory garden concepts, practitioners encountered challenges in the upkeep and utilization due to limited space, water shortages, inadequate plant care knowledge, and a lack of understanding of how to integrate sensory gardens into teaching practices. The study also served as a benchmark for potential similar projects in the township and led to further studies stemming from the original research. This project may be of interest to ECD practitioners and scholars, as well as education policymakers. Recommendations for future research, policy, and practice are offered.
Challenges in sexual and reproductive healthcare access for queer people in Gauteng, South Africa
(AOSIS, 2024-11) Seretlo, Raikane James; Smuts, Hanlie; Mokgatle, Mathildah Mpata
BACKGROUND: Sexual and reproductive healthcare services and needs (SRHS) are essential for improving overall health and well-being. Queer people face obstacles not only in obtaining and receiving SRHS but also in the provision of these services by healthcare professionals (HCPs). AIM: The study explored and described the challenges experienced by HCPs and queer people when providing and accessing SRHS in the Gauteng province. SETTING: We collected data from seven district public hospitals and two non-governmental organisation clinics that focussed on queer-related healthcare in the Gauteng province, South Africa. METHODS: This was an explorative-descriptive qualitative study, in which we conducted 55 one-on-one interviews. The researchers employed purposive sampling to select all HCPs, and respondent-driven sampling for all queer people who participated in the study. RESULTS: Six main themes emerged, demonstrating that HCPs and queer people faced similar, contrasting and differing challenges when rendering and receiving SRHS. These themes include HCPs’ belief that queer people are afraid, while queer people perceive HCPs as having negative attitudes and acting as gatekeepers. HCPs expressed surprise and confusion regarding gender identity, healthcare disparities and familial issues, which highlighted their feelings of incompetence in providing queer-related healthcare and their engagement with queer people as a barrier. CONCLUSION: Policies need to be strengthened to ensure inclusivity in the healthcare sector, thus, addressing SRHS for queer people. CONTRIBUTION: The findings from this study have the potential to bridge discrepancies and address the challenges faced by both HCPs and queer people through policy development.
Operationalizing a weighted performance scoring model for sustainable e-learning in medical education : insights from expert judgement
(Academic Publishing International Limited, 2024-07) Oluwadele, Deborah; Singh, Yashik; Adeliyi, Timothy; deborah.oluwadele@up.ac.za
Validation is needed for any newly developed model or framework because it requires several real-life applications. The investment made into e-learning in medical education is daunting, as is the expectation for a positive return on investment. The medical education domain requires data-wise implementation of e-learning as the debate continues about the fitness of e-learning in medical education. The domain seldom employs frameworks or models to evaluate students' performance in e-learning contexts. However, when utilized, the Kirkpatrick evaluation model is a common choice. This model has faced significant criticism for its failure to incorporate constructs that assess technology and its influence on learning. This paper aims to assess the efficiency of a model developed to determine the effectiveness of e-learning in medical education, specifically targeting student performance. The model was validated through Delphi-based Expert Judgement Techniques (EJT), and Cronbach's alpha was used to determine the reliability of the proposed model. Simple Correspondence Analysis (SCA) was used to measure if stability is reached among experts. Fourteen experts, professors, senior lecturers, and researchers with an average of 12 years of experience in designing and evaluating students' performance in e-learning in medical education participated in the evaluation of the model based on two rounds of questionnaires developed to operationalize the constructs of the model. During the first round, the model had 64 % agreement from all experts; however, 100% agreement was achieved after the second round, with all statements achieving an average of 52% strong agreement and 48% agreement from all 14 experts; the evaluation dimension had the most substantial agreements, next to the design dimension. The results suggest that the model is valid and may be applied as Key Performance Metrics when designing and evaluating e-learning courses in medical education.
The whole of cyber defense : syncing practice and theory
(Elsevier, 2024-12) De Nobrega, Kristel M.; Rutkowski, Anne-F.; Saunders, Carol S.
This review explores the problems Chief Information Security Officers (CISOs) and other cyber professionals face when defending their organization against cyberattacks. Using a Cyber Defense Grid, which was developed based on military science and cybersecurity concepts and terminology, we coded 125 articles published in Information Systems (IS) journals. We also employed three avenues (i.e., lenses) from cybernetic theory to frame the coding results to derive cyber defense strategies. In particular, we propose three strategic cyber defense modes: reactive, heuristic, and proactive. Taken together, these three modes suggest ways in which organizations can react strategically within the whole cyber defense domain.
Navigating the fourth industrial revolution : a systematic review of technology adoption model trends
(Emerald, 2024) Kruger, Sean; Steyn, Adriana Aletta (Riana); sean.kruger@up.ac.za
PURPOSE – Several disciplines and thousands of studies have used, developed and supported technology adoption theories to guide industry and support innovation. However, within the past decade, a paradigm shift referred to as the fourth industrial revolution (4IR) has resulted in new considerations affecting how models are used to guide emerging technology integration into business strategy. The purpose of this study is to determine which technology adoption model, or models are primarily used when assessing smart technologies in the 4IR construct. It is not to investigate the rigour of existing models or their theoretical underpinnings, as this has been proven. DESIGN/METHODOLOGY/APPROACH – To achieve this, a systematic literature review based on the preferred reporting items for systematic reviews and meta-analysis methodology is used. From 3,007 publications, 125 papers between 2015 and 2021 were deemed relevant for thematic analysis. FINDINGS – From the literature, five perspectives were extracted. As with other information and communication technology studies, the analysis confirms that the technology acceptance model remains the predominantly used model. However, 105 of the 125 models extended their theoretical underpinnings, indicating a lack of maturity. Furthermore, the countries of study and authors’ expertise are predominantly clustered in the European and Asian regions, despite the study noting expansion into 16 different subject areas, far beyond the smaller manufacturing scope of Industry 4.0. ORIGINALITY/VALUE – This study contributes theoretically by providing a baseline to develop a generalisable 4IR model grounded on existing acceptance trends identified. Practically, these insights demonstrate the current trends for strategists and policymakers to understand technology adoption within the 4IR to direct efforts that support innovation development, an increasingly crucial factor for survival in the digital age. Future research can investigate the additional constructs that were impactful while considering the level of research they were applied to.
Analysing the performance and interpretability of CNN-based architectures for plant nutrient deficiency identification
(MDPI, 2024-06) Mkhatshwa, Junior; Kavu, Tatenda; Daramola, Olawande; wande.daramola@up.ac.za
Early detection of plant nutrient deficiency is crucial for agricultural productivity. This study investigated the performance and interpretability of Convolutional Neural Networks (CNNs) for this task. Using the rice and banana datasets, we compared three CNN architectures (CNN, VGG16, Inception-V3). Inception-V3 achieved the highest accuracy (93% for rice and banana), but simpler models such as VGG-16 might be easier to understand. To address this trade-off, we employed Explainable AI (XAI) techniques (SHAP and Grad-CAM) to gain insights into model decision-making. This study emphasises the importance of both accuracy and interpretability in agricultural AI and demonstrates the value of XAI for building trust in these models.
A conceptual framework for human-centric and semantics-based explainable event detection
(Wiley, 2024-11) Kolajo, Taiwo; Daramola, Olawande; wande.daramola@up.ac.za
Explainability in the field of event detection is a new emerging research area. For practitioners and users alike, explainability is essential to ensuring that models are widely adopted and trusted. Several research efforts have focused on the efficacy and efficiency of event detection. However, a human-centric explanation approach to existing event detection solutions is still lacking. This paper presents an overview of a conceptual framework for human-centric semantic-based explainable event detection with the acronym HUSEED. The framework considered the affordances of XAI and semantics technologies for human-comprehensible explanations of events to facilitate 5W1H explanations (Who did what, when, where, why, and how). Providing this kind of explanation will lead to trustworthy, unambiguous, and transparent event detection models with a higher possibility of uptake by users in various domains of application. We illustrated the applicability of the proposed framework by using two use cases involving first story detection and fake news detection.

Browse

Recent Submissions