Department of Computer Sciences

Permanent URI for this collection

https://repository.run.edu.ng/handle/123456789/72

Browse

Now showing 1 - 12 of 12

Atomic Commit in Distributed Database Systems: The Approaches of Blocking and Non-Blocking Protocols
(International Journal of Engineering Research & Technology, 2014-10) Olowookere, Toluwase Ayobami
In distributed database systems, the primary need for commit protocols is to maintain the atomicity of distributed transactions. Atomic commitment issue is of prime importance in the distributed system and the issue becomes more necessary to deal with if some of the sites participating in the execution of the transaction commitment fail. Several atomic commit protocols have evolved to terminate distributed transactions. This paper presents an overview of a distributed transaction model, and a description of the two phase commit (2PC) protocol (which is blocking) and the one phase (1PC) commit protocols (which is non-blocking). This paper further examines the assumptions of these commit protocols in their bid to addressing the atomic commitment issue in distributed database systems. By restricting possible encountered failure to site failure, drawbacks in the assumptions of these atomic commit protocols were identified, which clearly show that the nonblocking protocol studied addresses the drawbacks of the widely used blocking protocol, 2PC, but in itself is no messiah (as it also constitutes drawbacks in practice). This work will spur other researchers to a more vigorous reconsideration of the 1PC nonblocking protocol.
Comparative Study and Detection of COVID-19 and Related Viral Pneumonia Using Fine-Tuned Deep Transfer Learning
(Springer- Intelligent Systems Reference Library, 2021) Olowookere, Toluwase Ayobami
Coronavirus (or COVID-19), which came into existence in 2019, is a viral pandemic that causes illness and death in the lives of human. Relentless research efforts have been on to improve key performance indicators for detection, isolation and early treatment. The aim of this study is to conduct a comparative study on the detection of COVID-19 and develop a Deep Transfer Learning Convolutional Neural Network (DTL-CNN) Model to classify chest X-ray images in a binary classification task (as either COVID-19 or Normal classes) and a three-class classification scenario (as either COVID-19, Viral-Pneumonia or Normal categories). Dataset was collected from Kaggle website containing a total of 600 images, out of which 375 were selected for model training, validation and testing (125 COVID-19, 125 Viral Pneumonia and 125 Normal). In order to ensure that the model generalizes well, data augmentation was performed by setting the random image rotation to 15 degrees clockwise. Two experiments were performed where a fine-tuned VGG-16 CNN and a fine-tuned VGG-19 CNN with Deep Transfer Learning (DTL) were implemented in Jupyter Notebook using Python programming language. The system was trained with sample datasets for the model to detect coronavirus in chest X-ray images. The fine-tuned VGG-16 and VGG-19 DTL models were trained for 40 epochs with batch size of 10, using Adam optimizer for weight updates and categorical cross entropy loss function. A learning rate of 1e−2 was used in fine-tuned VGG-16 while 1e−1 was used in fine-tuned VGG-19, and was evaluated on the 25% of the X-ray images. It was discovered that the validation and training losses were significantly high in the earlier epochs and then noticeably decreases as the training occurs in more subsequent epochs. Result showed that the fine-tuned VGG-16 and VGG-19 models, in this work, produced a classification accuracy of 99.00% for binary classes, and 97.33% and 89.33% for multi-class cases respectively. Hence, it was discovered that the VGG-16 based DTL model classified COVID-19 better than the VGG-19 based DTL model. Using the best performing fine-tuned VGG-16 DTL model, tests were carried out on 75 unlabeled images that did not participate in the model training and validation processes. The proposed models, in this work, provided accurate diagnostics for binary classification (COVID-19 and Normal) and multi-class classification (COVID-19,Viral Pneumonia and Normal), as it outperformed other existing models in the literature in terms of accuracy.
A Comparative Study of Two Microprocessor Based Distributed Systems: Intel Xeon and AMD Opteron
(IOSR Journal of Computer Engineering, 2014-09) Olowookere, Toluwase Ayobami
In this article, we draw a comparative study of microprocessor–based distributed systems, using the two major processors; Intel and AMD. Although the philosophy of their microarchitecture is the same, they differ in their approaches to implementation. Whether to increase the number of cores or to maximize the cores by hyperthreading, many of these features arise from philosophical grounds. These differences have been considered on the basis of their threading capabilities, coprocessor communications, memory accesses, and virtualization supports. . Moreover, the invention of hybrid server of these two processors cannot be compared with that of a server built with their individual processors. From our findings therefore, Intel remains the giant in microprocessor world while AMD on the other hand is in the front line of technological innovations.
The Design of a Hybrid Model-Based Journal Recommendation System
(Advances in Science, Technology and Engineering Systems Journal, 2020) Olowookere, Toluwase Ayobami
There is currently an overload of information on the internet, and this makes information search a challenging task. Researchers spend a lot of man-hour searching for journals related to their areas of research interest that can publish their research output on time. In, this study, a recommender system that can assist researchers access relevant journals that can publish their research output on time based on their preferences is developed. This system uses the information provided by researchers and previous authors' research publications to recommend journals with similar preferences. Data were collected from 867 respondents through an online questionnaire and from existing publication sources and databases on the web. The scope of the research was narrowed down to computer science related journals. A hybrid model-based recommendation approach that combined Content-Based and Collaborative filtering was employed for the study. The Naive Bayes and Random Forest algorithms were used to model the recommender. WEKA, a machine learning tool, was used to implement the system. The result of the study showed that the Naïve Bayes produced a shorter training time (0.01s) and testing time (0.02s) than the Random forest training time (0.41) and testing time (0.09). On the other hand, the classification accuracy of the Random forest algorithm outperformed the naïve Bayes with % correctly classified instance of 89.73 and 72.66; kappa of 0.893 and 0.714; True Positive of 0.897 and 0.727 and ROC area of 0.998 and 0.977, respectively, among other metrics. The model derived in this work was used as a knowledge-base for the development of a web-based application, named "Journal Recommender" which allowed academic authors to input their preferences and obtain prompt journal recommendations. The developed system would help researchers to efficiently choose suitable journals to help their publication quest.
Development of a Clinical Predictive Model for Stratification of Cancerous Diseases: A Case Study of Chronic Myeloid Leukemia
(International Journal of Advanced Science and Technology, 2020) Olowookere, Toluwase Ayobami
Scoring systems are typically used to stratify Chronic Myeloid Leukemia (CML) disease into their risk groups towards cure and survival prolongation. These systems, however, do not computationally handle very large datasets due to noise and overfitting of data. In literature, Machine Learning (ML) algorithms have been used to extract meaningful information from datasets, and their performances measured based on metrics such as accuracy and time to learn, among others. Nevertheless, the loss function (empirical risk) of the ML algorithms has been found not to have been largely considered to determine the risks incurred in adopting the ML algorithms for stratification. The aim of this study was to develop an Empirical Risk Minimization Data Stratification (ERMDS) algorithm that can aid the stratification of Chronic Myeloid Leukemia dataset. The algorithm developed would aid the development of a clinical predictive model using an application called ChroMyL app. A secondary dataset of 1640 CML patients, between 2003 and 2017 was collected from Obafemi Awolowo University Teaching Hospitals Complex, Ile-Ife, Osun State, Nigeria, and mined in WEKA 3.8.0 using basophil count and spleen size values on four ML algorithms (BayesNet, Multilayered perceptron, Projective Adaptive Resonance Theory (PART) and Logistic Regression). The algorithm with the highest performance was used in developing the ERMDS algorithm. Based on the analysis of the four classification algorithms carried out on five performance metrics which are: correctly classified instance, time to learn, kappa statistics, sensitivity and specificity, Logistic Regression had the highest accuracy value of 99.82%. As such, the ERMDS algorithm was developed using L1-regularized logistic regression solver in LibLINEAR 2.20. A Clinical Predictive Model (deployed as, ChroMyL app) was implemented with Javascript scripting language and jQuery on Macromedia Dreamweaver 16.0 to enhance page interactivity. The findings provided better insight into the process of adopting empirical risk minimization techniques in machine learning algorithms to solve disease risk group stratification problems, thus revealing how machine learning algorithms can be applied to real-world problems. The outcome of this study would provide more insight into the theoretical foundations of ML, and the important factors that must be put into consideration in every predictive or stratification models. Future researches can focus more on determining the loss function of other machine learning algorithms used in stratifying the chronic myeloid leukemia disease. Also, the approach to the design of the clinical predictive model application called ChroMyL app could be used for related cases.
A Framework for Detecting Credit Card Fraud with Cost-sensitive Meta-learning Ensemble Approach
(Elsevier- Scientific African, 2020) Olowookere, Toluwase Ayobami
Electronic payment systems continue to seamlessly aid business transactions across the world, and credit cards have emerged as a means of making payments in E-payment systems. Fraud due to credit card usage has, however, remained a major global threat to financial institutions with several reports and statistics laying bare the extent of this challenge. Several machine learning techniques and approaches have been established to mitigate this prevailing menace in payment systems, effective amongst which are ensemble methods and cost-sensitive learning techniques. This paper proposes a framework that combines the potentials of meta-learning ensemble techniques and cost-sensitive learning paradigm for fraud detection. The approach of the proposed framework is to allow base-classifiers to fit traditionally while the cost-sensitive learning is incorporated in the ensemble learning process to fit the cost-sensitive meta-classifier without having to enforce cost-sensitive learning on each of the base-classifiers. The predictive accuracy of the trained cost-sensitive meta-classifier and base classifiers were evaluated using Area Under the Receiver Operating Characteristic curve (AUC). Results obtained from classifying unseen data show that the cost-sensitive ensemble classifier maintains an excellent AUC value indicating consistent performance across different fraud rates in the dataset. These results indicate that the cost-sensitive ensemble framework is efficient in producing cost-sensitive ensemble classifiers that are capable of effectively detecting fraudulent transactions in different databases of payment systems irrespective of the proportion of fraud cases as compared to the performances of ordinary ensemble classifiers.
Hybridized Deep Convolutional Neural Network and Fuzzy Support Vector Machines for Breast Cancer Detection
(Springer Nature Computer Science, 2022) Olowookere, Toluwase Ayobami
A cancerous development that originates from breast tissue is known as breast cancer, and it is reported to be the leading cause of women death globally. Previous researches have proved that the application of Computer-Aided Detection (CADe) in screening mammography can assist the radiologist in avoiding missing breast cancer cases. However, many of the existing systems are prone to false detections or misclassifications and are majorly tailored towards either binary classification or three-class classification. Therefore, this study seeks to develop both two-class and three-class models for breast cancer detection and classification employing a deep convolutional neural network (DCNN) with fuzzy support vector machines. The models were developed using mammograms downloaded from the digital database for screening mammography (DDSM) and curated breast imaging subset CBISDDSM data repositories. The datasets were pre-processed, and features extracted for classification with DCNN and fuzzy support vector machines (SVM). The system was evaluated using accuracy, sensitivity, AUC, F1-score, and confusion matrix. The 3-class model gave an accuracy of 81.43% for the DCNN and 85.00% accuracy for the fuzzy SVM. The first layer of the serial 2-layer DCNN with fuzzy SVM for binary prediction yielded 99.61% and 100.00% accuracy, respectively. However, the second layer gave 86.60% and 91.65%, respectively. This study’s contribution to knowledge includes the hybridization of deep convolutional neural network with fuzzy support vector machines to improve the detection and classification of cancerous and non-cancerous breast tumours in both binary classification and three-class classification scenarios.
Immediate Word Recall in Cognitive Assessment Can Predict Dementia Using Machine Learning Techniques
(BMC, Alzheimer's Research & Therapy, 2023-06-15) Olowookere, Toluwase Ayobami
Background Dementia, one of the fastest-growing public health problems, is a cognitive disorder known to increase in prevalence as age increases. Several approaches had been used to predict dementia, especially in building machine learning (ML) models. However, previous research showed that most models developed had high accuracies, and they suffered from considerably low sensitivities. The authors discovered that the nature and the scope of the data used in this study had not been explored to predict dementia based on cognitive assessment using ML techniques. Therefore, we hypothesized that using word-recall cognitive features could help develop models for the prediction of dementia through ML techniques and emphasized assessing the models’ sensitivity performance. Methods Nine distinct experiments were conducted to determine which responses from either sample person (SP)’s or proxy’s responses in the “word-delay,” “tell-words-you-can-recall,” and “immediate-word-recall” tasks are essential in the prediction of dementia cases, and to what extent the combination of the SP’s or proxy’s responses can be helpful in the prediction of dementia. Four ML algorithms (K-nearest neighbors (KNN), decision tree, random forest, and artificial neural networks (ANN)) were used in all the experiments to build predictive models using data from the National Health and Aging Trends Study (NHATS). Results In the first scenario of experiments using “word-delay” cognitive assessment, the highest sensitivity (0.60) was obtained from combining the responses from both SP and proxies trained KNN, random forest, and ANN models. Also, in the second scenario of experiments using the “tell-words-you-can-recall” cognitive assessment, the highest sensitivity (0.60) was obtained by combining the responses from both SP and proxies trained KNN model. From the third set of experiments performed in this study on the use of “Word-recall” cognitive assessment, it was equally discovered that the use of combined responses from both SP and proxies trained models gave the highest sensitivity of 1.00 (as obtained from all the four models). Conclusion It can be concluded that the combination of responses in a word recall task as obtained from the SP and proxies in the dementia study (based on the NHATS dataset) is clinically useful in predicting dementia cases. Also, the use of “word-delay” and “tell-words-you-can-recall” cannot reliably predict dementia as they resulted in poor performances in all the developed models, as shown in all the experiments. However, immediate-word recall is reliable in predicting dementia, as seen in all the experiments. This, therefore, shows the significance of immediate-word-recall cognitive assessment in predicting dementia and the efficiency of combining responses from both SP and proxies in the immediate-word-recall task.
Modeling a Deep Transfer Learning Framework for the Classification of COVID-19 Radiology Dataset
(PeerJ Computer Science, 2021) Olowookere, Toluwase Ayobami
Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-Coronavirus-2 or SARS-CoV-2), which came into existence in 2019, is a viral pandemic that caused coronavirus disease 2019 (COVID-19) illnesses and death. Research showed that relentless efforts had been made to improve key performance indicators for detection, isolation, and early treatment. This paper used Deep Transfer Learning Model (DTL) for the classification of a real-life COVID-19 dataset of chest X-ray images in both binary (COVID-19 or Normal) and three-class COVID-19, Viral-Pneumonia or Normal) classification scenarios. Four experiments were performed where fine-tuned VGG-16 and VGG-19 Convolutional Neural Networks (CNNs) with DTL were trained on both binary and three-class datasets that contain X-ray images. The system was trained with an X-ray image dataset for the detection of COVID-19. The fine-tuned VGG-16 and VGG-19 DTL were modelled by employing a batch size of 10 in 40 epochs, Adam optimizer for weight updates, and categorical cross-entropy loss function. The results showed that the fine-tuned VGG-16 and VGG-19 models produced an accuracy of 99.23% and 98.00%, respectively, in the binary task. In contrast, in the multiclass (three-class) task, the fine-tuned VGG-16 and VGG-19 DTL models produced an accuracy of 93.85% and 92.92%, respectively. Moreover, the fine-tuned VGG-16 and VGG-19 models have MCC of 0.98 and 0.96 respectively in the binary classification, and 0.91 and 0.89 for multiclass classification. These results showed strong positive correlations between the models’ predictions and the true labels. In the two classification tasks (binary and three-class), it was observed that the fine-tuned VGG-16 DTL model had stronger positive correlations in the MCC metric than the fine-tuned VGG-19 DTL model. The VGG-16 DTL model has a Kappa value of 0.98 as against 0.96 for the VGG-19 DTL model in the binary classification task, while in the three-class classification problem, the VGG-16 DTL model has a Kappa value of 0.91 as against 0.89 for the VGG-19 DTL model. This result is in agreement with the trend observed in the MCC metric. Hence, it was discovered that the VGG-16 based DTL model classified COVID-19 better than the VGG-19 based DTL model. Using the best performing fine-tuned VGG-16 DTL model, tests were carried out on 470 unlabeled image dataset, which was not used in the model training and validation processes. The test accuracy obtained for the model was 98%. The proposed models provided accurate diagnostics for both the binary and multiclass classifications, outperforming other existing models in the literature in terms of accuracy, as shown in this work.
Performance Evaluation of an Ensemble Method for Diagnosis of Chronic Kidney Disease with Feature Selection Technique
(IEEE Xplore: 2020 International Conference on Decision Aid Sciences and Application (DASA), 2020) Olowookere, Toluwase Ayobami
Chronic Kidney Disease (CKD) is a public health issue which is seen as a significant threat to human life due to abnormal functioning of kidney over a period of months or years which, if left untreated, may damage vital organs in the body leading to an increased rate in cardiovascular mortality which may result in sudden death if not early detected. Data mining techniques are employed in several clinical diagnoses for making intelligent diagnostics decisions that can be applied in disease prediction. The performances of these techniques are very promising in the management of different ailments to reduce the high numbers of people that die yearly due to inaccurate diagnosis of numerous disease conditions. This study evaluates the performance of a bagging ensemble technique on CKD dataset with an effective feature selection technique to yield a reliable and accurate predictive model capable of correctly classifying diseased from non-diseased patients. The study was investigated on a real patient dataset obtained from the UCI machine learning repository consisting of 400 instances with 24 conditional attributes and a decisional class. Radom forest algorithm was used as a measure to select the best subset of features for the predictive models. Naïve Bayes, k-Nearest Neighbor, and Decision Tree algorithms serve as the base classifiers whose performance were aggregated using the bagging ensemble approach to improve base learners' performances. Results obtained from the study showed the effect of feature selection and ensemble technique in improving the accuracy of data mining classification algorithms. The model's optimal result is achieved using 7 best-selected features on the ensemble classifier with 100% accuracy of CKD diagnosis compared to 98.3% accuracy without feature selection. Hence, making the model suitable for efficient diagnosis of CKD.
A Topic Modelling-Based Framework for Mining Digital Library’s Text Documents
(African Journal of Computing & ICT, 2015-12) Olowookere, Toluwase Ayobami
The impacts and contributions of scholarly research work in the economic growth and sustainability of any nation cannot be overemphasized. The digital library has emerged as a reliable resource for provisioning researchers with scholarly knowledge (erudition that result from the research works) which are documented and published in form of journal articles, technical reports or conference proceedings amongst others. However, as academic institutions and publishers around the world are choosing to make their thesis, dissertation and journal articles available in digital form, this electronic repository of knowledge (the digital library), though organized, is flooded with an exploding large collections of documents filled with hidden but useful information in form of the varieties of topics of discourse inherent in them. Thus making it imperative to develop a flexible means to automatically discover the topics that pervade the collections in such digital library. Currently, the application of topic modelling technique holds great promises and has tremendous results in extracting the topical contents of document corporal. In this regard, this paper presents a Topic Modelling-based framework for mining document collections of a digital library for topical structure discovery alongside topic-based similarities search between document collection pairs, by means of integrating the base topic modelling algorithm and inverted Kullback-Leibler divergence mechanism. The framework shows potency in the automatic discovery of topical structures of document collections and it as well describes the capability of finding topic-based similarities between document collection pairs.
UPH Digital Library Miner: A Topic Modelling-based Software Application for Mining Document Collections of a Digital Library
(International Journal of Computer Applications, 2015-12) Olowookere, Toluwase Ayobami
With changing user expectations, many traditional libraries are moving toward digital content storage. Accessible from anywhere at any time, digital contents as stored in digital libraries provide users with efficient, on-demand information experiences. With this trend, the amount of digital contents especially digital text documents made available to users have tremendously increased over the years, being filled with hidden information in form of the varieties of topics of discourse inherent in them leading to information overload. Accordingly, users, mostly computational researchers are presented with challenges on the discovery and identification of the varieties of topical contents of the collections in the digital library thus making it imperative to develop a means to automatically discover the topics that pervade the collections in a digital library. This paper therefore presents UPH Digital Library Miner, a software application for mining document collections of a digital library for topical structure discovery and topic-based similarities search between collection pairs, using topic modeling algorithm and inverted Kullback-Leibler divergence measure. The application is integrated with document collections built in a widely used digital library software system— Greenstone digital library system, via loose-coupling integration approach. Results obtained from using this software application on the Greenstone’s document collections that contain abstracts of about 628 documents from IEEE transactions on Software Engineering show its ability to discover latent topical structures in collections and also report collections that are similar based on their discovered topical structure.

Browse

Browsing Department of Computer Sciences by Author "Olowookere, Toluwase Ayobami"

Results Per Page

Sort Options