DSpace Angular :: Browsing by Author "Babatunde, Ronke Seyi"

Browsing by Author "Babatunde, Ronke Seyi"

Now showing 1 - 9 of 9

A Hybridized Feature Extraction Model for Offline Yorùbá Document Recognition
(Asian Journal of Research in Computer Science, 2023-05-16) Ajao, Jumoke Falilat; Isiaka, Mope Rafiu; Babatunde, Ronke Seyi
Document recognition is required to convert handwritten and text documents into digital equivalents, making them more easily accessible and convenient to store. This study combined feature extraction techniques for recognizing Yorùbá documents in an effort to preserve the cultural values and heritages of the Yorùbá people. Ten Yorùbá documents were acquired from Kwara State University’s Library, and ten indigenous literate writers wrote the handwritten version of the documents. These were digitized using HP Scanjet300 and pre-processed. The pre-processed image served as input to the Local Binary Pattern, Speeded-Up-Robust-Features and Histogram of Gradient. The combined extracted feature vectors were input into the Genetic Algorithm. The reduced feature vector was fed into Support Vector Machine. A 10-folds cross-validation was used to train the model: LBP-GA, SURF-GA, HOG-GA, LBP-SURF-GA, HOG-SURF-GA, LBP-HOG-GA and LBP-HOG-SURF-GA. LBP-HOG-SURF-GA for Yorùbá printed text gave 90.0% precision, 90.3% accuracy and 15.5% FPR. LBP-HOG-SURF-GA for Handwritten Yorùbá document showed 80.9% precision, 82.6% accuracy and 20.4% (FPR) LBP-HOG-SURF-GA for CEDAR gave 98.0% precision, 98.4% accuracy and 2.6% FPR. LBP-HOG-SURF-GA for MNIST gave 99% precision, 99.5% accuracy, 99.0% and 1.1% FPR. The results of the hybridized feature extractions (LBP HOG-SURF) demonstrated that the proposed work improves significantly on the various classification metrics.
A Machine Learning Approach to Dropout Early Warning System Modeling
(International Journal of Advanced Studies in Computer Science and Engineering (Ijascse.）, 2019-02-28) Isiaka, Mope Rafiu; Babatunde, Ronke Seyi; Abdulsalam, Olaniyi Sulaiman
Dropout has been identified as a social problem with an immediate and long term consequences on the socio economic development of the society. A major element of a conceptual framework for dropout monitoring and control system is the prevention component. The component specified the need for automatic dropout early warning system. This paper presents the procedure for building the adaptive model, that is the core element for the realization of the prevention component of the framework. The model development is guided by the knowledge of the domain experts. An experimental approach was used to identify the K-Nearest Neighbors (KNN) as the best of the six algorithms considered for the adaptive models. Other algorithms explored are Logistic Regression (LR), Linear Discriminant Analysis (LDA), Classification and Regression Trees (CART), Gaussian Naive Bayes (NB) and Support Vector Machines (SVM). Historical dataset from a State University in the North Central region of Nigeria was used for building and testing the adaptive model. The average performance of the model on full attributes and the first year academic attributes compared favorably with 0.98 Precision, Recall and F1-Scores. Conversely, the Precision, Recall and F1-Scores on the entry attributes are 0.90, 0.95 and 0.92 respectively. The implication of this is that the academic performance is a very significant factor for dropping out of the educational system. The developed model can provide an early signal on the students with the propensity to dropout, thereby serving as an advisory system for the students, parents and school management towards curtailing the menace.
A Study of Impacts of Artificial Intelligence on COVID-19 Prediction, Diagnosis, Treatment, and Prognosis
(Journal of Advances in Mathematical & Computational Sciences, 2022-12-04) Isiaka, Rafiu Mope; Babatunde, Ronke Seyi; Ajao, Jumoke Falilat; Yusuff, Shakirat Ronke; Popoola, Damilola David; Arowolo, Michael Olaolu; Adewole, Kayode
Following the identification of Coronavirus Disease 2019 (COVID-19) in Wuhan, China in December 2019, AI researchers have teamed up with a health specialist to combat the virus. This study explores the medical and non-medical areas of COVID-19 that AI has impacted: the prevalence of the AI technologies adopted across all stages of the pandemic, the collaboration networks of global AI researchers, and the open issues. 21,219 papers from ACM Digital, Science Direct and Google Scholar were examined. Adherence to the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) framework and utilizing the PICO (population, intervention, comparison, and outcome) paradigm, guided the inclusion of researches in the review. Tables and graphs were utilized to display the results. Analysis revealed that AI has impacted 4 molecular, 4 clinical, and 7 societal areas of COVID-19. Deep Learning among other AI technologies was traced to all aspects of the pandemic. 2173 authors and co-authors were traced to these achievements, while 32 of the most connected 51 authors were affiliated with institutions in China, 18 to the United States, and 1 to Europe. The open issues identified had to do with the quality of datasets, AI model deployment, and privacy issues. This study demonstrates how AI may be utilized for COVID-19 diagnosis, prediction, medication and vaccine identification, prognosis, and contact person monitoring. This investigation began at the beginning of the epidemic and continued until the first batch of vaccinations received approval. The study provided collaboration opportunities for AI researchers and revealed open issues that will spike further research toward preparing the world for any future pandemic.
An Improved Coronary Heart Disease Predictive System Using Random Forest
(Asian Journal of Research in Computer Science, 0021-08-11) Abdulraheem Abdul; Isiaka, Mope Rafiu; Babatunde, Ronke Seyi; Ajao, Jumoke Falilat
Aims: This work aim is to develop an enhanced predictive system for Coronary Heart Disease (CHD). Study Design: Synthetic Minority Oversampling Technique and Random Forest. Methodology: The Framingham heart disease dataset was used, which was collected from a study in Framingham, Massachusetts, the data was cleaned, normalized, rebalanced. Classifiers such as random forest, artificial neural network, naïve bayes, logistic regression, k-nearest neighbor and support vector machine were used for classification. Results: Random Forest outperformed other classifiers with an accuracy of 98%, a sensitivity of 99% and a precision of 95.8%. Feature selection was employed for better classification, but no significant improvement was recorded on the performance of the classifier with feature selection. Train test split also performed better that cross validation. Conclusion: Random Forest is recommended for research in Coronary Heart Disease prediction domain.
Beyondthehypesondatarebalancinginimbalancelearning: towards abalancedframeworkandrecommendersystem
(International Journal of Data Science and Analytics, 2025-07-21) Isiaka, Mope Rafiu; Olorede, Opeyemi Kabir; Babatunde, Ronke Seyi; Ajao, Jumoke Falilat
Class-imbalanced learning presents critical challenges in machine learning, largely because data in most domains are naturally imbalanced. Although resampling techniques have been widely applied to address this issue, their effectiveness has been inconsistent and sometimes flawed, owing to artificial assumptions. In this study, we move beyond hype surrounding resampling methods by exploring alternative strategies, such as ensemble learning, cost-sensitive algorithms, and one-class classification techniques. Through rigorous experimentation across extreme, moderate, and mild imbalance levels, our findings reveal that these alternatives often outperform traditional resampling in terms of F1-scores, with ensemble SVM and one class logistic regression achieving notable values of 1.00 and 0.90, respectively. In addition, we introduce a knowledge-based recommender system designed to help practitioners choose the most appropriate techniques for addressing class imbalance. This research argues that resampling is not always the optimal solution for all instances, thereby advocating a more balanced framework that leverages advanced methods for superior performance in imbalanced learning tasks. Our study advances the field by offering a pragmatic, data-driven approach to overcoming class imbalances, contributing valuable insights for both researchers and practitioners.
Development of a Web-Based Platform Utilizing a Hybrid CNN Architecture for Real-Time Fecal Image Classification
(Ajayi Crowther Journal of Pure and Applied Sciences, 2025-10-29) Omideyi, Damilare Andrew; Isiaka, Mope Rafiu; Babatunde, Ronke Seyi
Accurate classification of fecal images offers a non-invasive and scalable approach to detecting enteric diseases such as Coccidiosis, Salmonellosis, and Newcastle disease. Traditional diagnostic methods remain limited by their dependence on expert analysis, time consumption, and cost. This study presents the development of a web-based platform utilizing a hybrid Convolutional Neural Network (CNN) architecture for the real-time classification of fecal images. The proposed model combines MobileNetV2 for lightweight inference and VGG-16 for deep feature extraction, ensuring both accuracy and computational efficiency. Following training on a curated and augmented dataset of labeled fecal images, the model was quantized and converted into TensorFlow Lite format for mobile compatibility. The final system enables users to upload images via a browser or Android application and receive instant classification into four categories: Coccidiosis, Salmonella, Newcastle disease, or Healthy. Performance evaluations confirm high diagnostic accuracy and responsiveness under real-world conditions. This study demonstrates that hybrid deep learning models, when integrated into web-based platforms, can support effective classification of fecal images for field-ready disease detection. It is recommended that future work expands the dataset across multiple geographic regions and incorporates additional diagnostic inputs to enhance system robustness and adaptability.
Implication of Variation in Ant Colony Optimization Algorithm (ACO) Parameters on Feature Subset Size
(Computing and Information Systems Journal, 2017) Babatunde, Ronke Seyi; Isiaka, Mope Rafiu; Oyeranmi J. Adigun; Babatunde, N. Akinbowale
Purpose: The purpose of this research is to apply ant colony optimization (ACO), a nature inspired computational (NIC) technique to achieve optimal feature subset selection, for the purpose of data dimensionality reduction, which will remove redundancy, provide a reduced storage space, improve memory utilization and subsequently, classification task. Design/Methodology/Approach: The ACO feature subset selection process was filter-based, accomplished by using correlation coefficient as the heuristic information which guides the search to determine the pixel (feature) attractiveness. The intensity of the pixel represents the pheromone level . Also, ρ, φ, and are key parameters which were tuned to various arbitrary values at each iteration to determine the best combination of parameters which can result in a reduced feature subset (window size). The procedure was simulated in MATLAB 2012. Findings: The study has been able to demonstrate that experimental results obtained by tuning the parameters of ACO at each run gave a reduced feature subset (window size), at parameter settings of α (0.10), β (0.45), ρ (0.50), φ (1.0) which was salient for efficient face recognition thereby demonstrating the effectiveness and superiority of filter-based over wrapper–based feature selection. Research Limitation: The major limiting factor in this research was time and materials which hindered the further testing and implementation of innovative computational intelligence. Originality/Value: This research work is valuable because it gives the implication of the variation of the ACO parameters which determines the extent to which the optimal solution construction that results in a reduced feature subset can be realized.
Towards adoption of information and communication technology in higher education – a structural equation model approach
(International Journal of Learning Technology, 2021-06-02) Isiaka, Rafiu Mope; Babatunde, Ronke Seyi; Eiriemiokhale, Kennedy Arebamen; Popoola, Damilola David
Use of White Shark Optimization for Improving the Performance of Convolution Neural Network in Classification of Infected Citrus
(University of Ibadan Journal of Science and Logics in ICT Research, 2023-06-09) Babatunde, Ronke Seyi; Isiaka, Mope Isiaka; Ajao, Jumoke Falilat; Abdulsalam, Olaniyi Sulaiman; Balogun, Bukola Fatimah
Citrus plant diseases are major causes of reduction in the production of citrus fruits and their usage. Early detection of the onset of the diseases is very important to curb and reduce its spread. A number of researches have been done on the detection and classification of the diseases, most of which have identified poor labelling of symptoms which result into improper classification. Some researchers have also experimented on the effectiveness of convolution neural network and other deep learning techniques, most of which results into a faster convergence but suffers from low accuracy, computational overhead and overfitting as fundamental issues. To reduce the effects of overfitting, this research developed a White Shark Optimization-Convolution Neural Network (WSO-CNN) technique to address the aforementioned problem by introducing a regularization strategy via feature selection which selects more useful and distinguishing features for classification. As a result, the developed technique was able to detect and classify various types of citrus fruit diseases and label them accordingly with low false positive rate, high sensitivity, specificity, increased accuracy and reduced recognition time, based on all the experiments performed with the dataset used in the research. Hence WSOCNN performed better than CNN in classifying citrus plant disease having a reduced FPR of 3.57%, 8.34%, 3.89% and 9.00% for black spot, greasy spot, canker and healthy/non healthy dataset respectively.

Browsing by Author "Babatunde, Ronke Seyi"

Results Per Page

Sort Options