Department of Computer Sciences

Permanent URI for this collection

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

Browse

Recent Submissions

Now showing 1 - 5 of 120
  • No Thumbnail Available
    Item
    Efficient diagnosis of diabetes mellitus using an improved ensemble method
    (Scientific Reports, 2025-01) Olorunfemi, Blessing O.
    Diabetes is a growing health concern in developing countries, causing considerable mortality rates. While machine learning (ML) approaches have been widely used to improve early detection and treatment, several studies have shown low classification accuracies due to overfitting, underfitting, and data noise. This research employs parallel and sequential ensemble ML approaches paired with feature selection techniques to boost classification accuracy. The Pima India Diabetes Data from the UCI ML Repository served as the dataset. Data preprocessing included cleaning the dataset by replacing missing values with column means and selecting highly correlated features using forward and backward selection methods. The dataset was split into two parts: training (70%), and testing (30%). Python was used for classification in Jupyter Notebook, and there were two design phases. The first phase utilized J48, Classification and Regression Tree (CART), and Decision Stump (DS) to create a random forest model. The second phase employed the same algorithms alongside sequential ensemble methods—XG Boost, AdaBoostM1, and Gradient Boosting—using an average voting algorithm for binary classification. Evaluation revealed that XG Boost, AdaBoostM1, and Gradient Boosting achieved classification accuracies of 100%, with performance metrics including F1 score, MCC, Precision, Recall, AUC-ROC, and AUC-PR all equal to 1.00, indicating reliable predictions of diabetes presence. Researchers and practitioners can leverage the predictive model developed in this work to make quick predictions of diabetes mellitus, which could save many lives.
  • No Thumbnail Available
    Item
    APerformance Study of Selected Machine Learning Techniques for Predicting Heart Diseases
    (Springer, 2025-04) Olorunfemi, Blessing O.
    Heart Disease remains a leading cause of mortality worldwide. It alarmingly rises at a quick rate, making early heart disease prediction crucial for effective prevention and timely intervention. Heart disease diagnosis is a difficult process that requires technical skills and accuracy to complete. With improvements in technology, computing has lent its voice to simplify the diagnosis of various health problems. Machine learning uses past or existing history to predict future results. Various machine learning techniques have been developed over the years and used in predicting heart diseases with various levels of performance. Identifying the best-suited machine learning technique to use for prediction purposes can be a challenging task. This research work analyses the performance of seven (7) machine learning techniques, comprising AdaBoost Algorithm, KNN, Logistic Regression, Naïve Bayes Classifier, Random Forest, SVM, and XGBoost. The heart disease dataset was downloaded from the UCI repository and analysed using Python programming language in the Jupyter Notebook environment. A comparative analysis of the seven (7) techniques was performed based on Accuracy, Precision, and Recall. From the results obtained, KNN, Random Forest, and XGBoost showed superior performance over the others with an accuracy of 100%, AdaBoost Algorithm followed with an accuracy of 92.2%, SVM followed with an accuracy of 91.71%, Naïve Bayes Classifier followed with an accuracy of 88.29% while Logistic Regression has the least accuracy of 86.34%. KNN, RF, and XGBoost outperformed AdaBoost, SVN, and LR
  • No Thumbnail Available
    Item
    Development Of A Machine Learning Model For Brand And Audience Segmentation Using Demographic Data
    (Corpus Intellectual, 2025) Gbeminiyi Falowo
    The expansion of the global business landscape, a highimpact factor in eCommerce, has resulted in identifying potential customers and their positive reactions to products or services offered by companies that use the internet to promote their electronic business. With a high increase in audience using social media, there is a need for brand and audience segmentation and targeting for profit-making; thus, this study developed a machine learning model for brand and audience segmentation using the Social Media Advertising Dataset. The dataset includes comprehensive data on social media advertising campaigns across Facebook, Instagram, Pinterest, and Twitter, featuring ad impressions, clicks, spending, demographic targeting, and conversion rates. With 16 columns and 300,000 rows, the dataset offered substantial data for analysis. The study compared the performance of a Naive Bayes model with a Random Forest algorithm in two existing literature; the Naive Bayes model achieved an accuracy of 35%, the Random Forest model achieved an accuracy of 89.6%, and the Random Forest model in the current study's model reached 97% accuracy. The Random Forest model's superior performance in both studies demonstrates its effectiveness in consumer group segmentation, indicating its practical utility in optimizing marketing strategies and improving customer targeting. An implementation of the developed model of the study was in Python and deployed on a website using the Flask framework, providing an accessible tool for practical applications.
  • No Thumbnail Available
    Item
    Latency-Aware Load Distribution Model for Vehicular Adhoc Networks (VANETS)
    (International Journal of Innovative Science and Research Technology, 2022-03) Gbeminiyi Falowo
    Vehicular Adhoc Network (VANET) is a network which prominent and fluid features have helped in drawing contiguous attention by researchers for more than thirty years. More often than none, routing and securing the network are of utmost priority to researchers while little focus is given to ensuring the effective distribution of load in Vehicular communications in order to ensure an infinitesimal experience of latency. However, with the advent of Intelligent Transport System (ITS) in pair with the possibility of cloud and Internet of Things (IoT) in VANET that support a mickle spectrum of mobile distributed software, there is need for a system that will evenly distribute load in the network. This study thereby introduces a latency-aware model via a 3-tier load distribution mechanism that reduces delay in message transmission and also helps in addressing traffic congestion faster.
  • No Thumbnail Available
    Item
    Blockchain Mechanism Approach to Smothering of Denial of Service (DoS) Spikes: A Focus on Internet of Things (IoT) Technologies
    (International Journal of Research and Scientific Innovation, 2024-09) Gbeminiyi Falowo
    Denial of Service (DoS) is a cybercrime that attempts to impede electronic consumers from accessing websites and online services by saturating a server with internet traffic. Cyber-spikers use a network of infected computers, tools like bots, and other machines they can access remotely. A decade ago, businesses and financial institutions lost approximately half a trillion dollars due to DOS spikes. DoS savages would triple in number before the closure of the year 2023 from about eight million less than five years ago. This study uses a blockchain-based decentralized authentication technique to guard against DoS attacks on the application layer of Internet of Things (IoT) technologies. This secured mechanism involves starting the communication process, developing the system, and suggesting an intelligent contract. Performance evaluation of the developed model was carried out by comparing the approaches’ temporal complexity. The recommended method was also used on two processors operating at two distinct speeds while utilizing the SolarWinds application, an online CPU stress test, and usage with a deduction that the second is preferred. An Intelligent contract for IoT machine usage is established to authorize the blockchain level.