Femi Abiodun Elegbeleye
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Femi Abiodun Elegbeleye has a total of 10 co-authors. Here are the top co-authors by number of shared publications:
- Omobayo Ayokunle Esan
Walter Sisulu University
4 shared publications - Dorcas Oladayo Esan
Tshwane University of Technology
2 shared publications - Munienge Mbodila
Walter Sisulu University
2 shared publications - Chuma Mbaleki
Walter Sisulu University
1 shared publication - Ife Elegbeleye
Department of Physics, University of Pretoria, Hatfield, South Africa
1 shared publication - Kesewaa Koranteng
Walter Sisulu University
1 shared publication
Latest Publications
- Open Access
- 1 August 2025
- Published by Institute of Advanced Engineering and Science in Bulletin of Electrical Engineering and Informatics
- Vol. 14 (4) , 2721-2734
Prediction of stock market price for investors using machine learning approach
Stock market price prediction is a challenging task that plays a crucial role in investment decision-making and financial risk management. Traditional approaches often rely on a single machine learning (ML) algorithm for predictive modeling. In this contribution, an innovative framework that integrates logistic regression (LR) with support vector machine (SVM) to improve the accuracy and reliability of stock market price prediction. Combining the strengths of both algorithms, the proposed model harnesses the interpretability of LR and the robustness of SVM to capture complex relationships in stock market data. Experiments conducted on publicly available Yahoo Finance stock dataset and the Dhaka dataset, the results show that the proposed model yielded accuracies of 97.15% and 98.86% respectively. In comparison with other models, the proposed method outperformed the other models in terms of root mean square error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), mean squared error (MSE), and accuracy. The contribution and importance of leveraging hybrid modelling techniques to enhance stock market price prediction and facilitate informed investment decision-making.
- 1 August 2025
- Published by Institute of Advanced Engineering and Science in Bulletin of Electrical Engineering and Informatics
- Vol. 14 (4) , 2721-2734
Open AccessPrediction of stock market price for investors using machine learning approach
Stock market price prediction is a challenging task that plays a crucial role in investment decision-making and financial risk management. Traditional approaches often rely on a single machine learning (ML) algorithm for predictive modeling. In this contribution, an innovative framework that integrates logistic regression (LR) with support vector machine (SVM) to improve the accuracy and reliability of stock market price prediction. Combining the strengths of both algorithms, the proposed model harnesses the interpretability of LR and the robustness of SVM to capture complex relationships in stock market data. Experiments conducted on publicly available Yahoo Finance stock dataset and the Dhaka dataset, the results show that the proposed model yielded accuracies of 97.15% and 98.86% respectively. In comparison with other models, the proposed method outperformed the other models in terms of root mean square error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), mean squared error (MSE), and accuracy. The contribution and importance of leveraging hybrid modelling techniques to enhance stock market price prediction and facilitate informed investment decision-making.
- Open Access
- 1 March 2024
- Published by Institute of Advanced Engineering and Science in IAES International Journal of Artificial Intelligence (IJ-AI)
- Vol. 13 (1) , 247-255
Cost-effective internet of things privacy-aware data storage and real-time analysis
It has been estimated that about 20 billion internet of things (IoT) devices are currently connected to the Internet. This has led to voluminous data generation which makes storaging, managing, and decision making on data to be challenging. Hence, exposes users’ privacy to be vulnerable to unauthorized people. To address these issues, this research proposed cost-effective storage for keeping and processing the IoT data in real-time. The proposed Fframework utilized a reliable hybridised data privacy model to protect the personal information of users. An empirically evaluation was done to identify the best models using data k-anonymity (KA), l-diversity (LD), t-closeness (TC), and differential privacy (DP). The performance evaluation of cloud computing and fog computing was done through simulations. The results obtained show that the combination of two data privacy models: differential privacy and k-anonymity models performed better than any individual model and any other combined models in the protection of users’ personal information. Lastly, fog computing was found to perform better than the cloud in terms of latency, energy consumption, network usage and execution time. In conclusion, the current study strongly recommends the use of hybridised privacy model of differential privacy (DP) and k-anonymity (KA) for the protection of IoT generated data privacy.
- 1 March 2024
- Published by Institute of Advanced Engineering and Science in IAES International Journal of Artificial Intelligence (IJ-AI)
- Vol. 13 (1) , 247-255
Open AccessCost-effective internet of things privacy-aware data storage and real-time analysis
It has been estimated that about 20 billion internet of things (IoT) devices are currently connected to the Internet. This has led to voluminous data generation which makes storaging, managing, and decision making on data to be challenging. Hence, exposes users’ privacy to be vulnerable to unauthorized people. To address these issues, this research proposed cost-effective storage for keeping and processing the IoT data in real-time. The proposed Fframework utilized a reliable hybridised data privacy model to protect the personal information of users. An empirically evaluation was done to identify the best models using data k-anonymity (KA), l-diversity (LD), t-closeness (TC), and differential privacy (DP). The performance evaluation of cloud computing and fog computing was done through simulations. The results obtained show that the combination of two data privacy models: differential privacy and k-anonymity models performed better than any individual model and any other combined models in the protection of users’ personal information. Lastly, fog computing was found to perform better than the cloud in terms of latency, energy consumption, network usage and execution time. In conclusion, the current study strongly recommends the use of hybridised privacy model of differential privacy (DP) and k-anonymity (KA) for the protection of IoT generated data privacy.
- Open Access
- 30 November 2023
- Published by OU Scientific Route in EUREKA: Social and Humanities
- No. 6,p. 3-12
Power supply rationing in an era of e-learning: evidence from the rural university
In a world, driven by technology, South Africa's battle with power supply rationing, commonly known as load shedding, is sending shockwaves through the education landscape. This study investigates the intricate relationship between load shedding and e-learning, focusing on a rural university in South Africa. It explores how the disruption of the power supply affects the quest for digital education and sheds light on the hidden challenges, faced by both students and lecturers. The energy crisis loomed large, posing a unique threat to the country's academic advancement. This study was carried out at Walter Sisulu University, meticulously dissects the impacts of load shedding on e-learning, providing valuable insights into how these unforeseen challenges are altering the educational landscape. The research blends quantitative data analysis with a qualitative perspective, capturing the real-life experiences of educators and learners navigating a world where power supply can change the course of a digital classroom. The research elucidated the dominance of e-learning during lockdown regulations to curb the spread of coronavirus and a hybrid learning pedagogy that combines both e-learning and physical classes in the post-covid era. The research revealed that e-learning is significantly affected by Loadshedding.
- 30 November 2023
- Published by OU Scientific Route in EUREKA: Social and Humanities
- No. 6,p. 3-12
Open AccessPower supply rationing in an era of e-learning: evidence from the rural university
In a world, driven by technology, South Africa's battle with power supply rationing, commonly known as load shedding, is sending shockwaves through the education landscape. This study investigates the intricate relationship between load shedding and e-learning, focusing on a rural university in South Africa. It explores how the disruption of the power supply affects the quest for digital education and sheds light on the hidden challenges, faced by both students and lecturers. The energy crisis loomed large, posing a unique threat to the country's academic advancement. This study was carried out at Walter Sisulu University, meticulously dissects the impacts of load shedding on e-learning, providing valuable insights into how these unforeseen challenges are altering the educational landscape. The research blends quantitative data analysis with a qualitative perspective, capturing the real-life experiences of educators and learners navigating a world where power supply can change the course of a digital classroom. The research elucidated the dominance of e-learning during lockdown regulations to curb the spread of coronavirus and a hybrid learning pedagogy that combines both e-learning and physical classes in the post-covid era. The research revealed that e-learning is significantly affected by Loadshedding.
- Open Access
- 1 June 2023
- Published by Institute of Advanced Engineering and Science in Bulletin of Electrical Engineering and Informatics
- Vol. 12 (3) , 1674-1683
Surveillance detection of anomalous activities with optimized deep learning technique in crowded scenes
The performance of conventional surveillance systems is challenged by high error detection rates in busy scenes, which has significantly affected the accurate detection of the current surveillance system. Feature representation and object pattern extraction from different scenes have made deep learning (DL) promising methods in surveillance systems, compared to the approaches where features are created manually. To improve the detection accuracy, this paper presents an intelligent DL technique that combines convolutional neural network (CNN) and long short-term memory (LSTM). CNN extracts and learns the object features from a set of raw images, while the LSTM is then used by gated mechanisms to store important information from the extracted features. The proposed method was validated using datasets from the University of California San Diego (UCSD). The result shows that the model achieves 95% accuracy, which is superior compared to other conventional detection models.
- 1 June 2023
- Published by Institute of Advanced Engineering and Science in Bulletin of Electrical Engineering and Informatics
- Vol. 12 (3) , 1674-1683
Open AccessSurveillance detection of anomalous activities with optimized deep learning technique in crowded scenes
The performance of conventional surveillance systems is challenged by high error detection rates in busy scenes, which has significantly affected the accurate detection of the current surveillance system. Feature representation and object pattern extraction from different scenes have made deep learning (DL) promising methods in surveillance systems, compared to the approaches where features are created manually. To improve the detection accuracy, this paper presents an intelligent DL technique that combines convolutional neural network (CNN) and long short-term memory (LSTM). CNN extracts and learns the object features from a set of raw images, while the LSTM is then used by gated mechanisms to store important information from the extracted features. The proposed method was validated using datasets from the University of California San Diego (UCSD). The result shows that the model achieves 95% accuracy, which is superior compared to other conventional detection models.