Covenant Journal of Informatics and Communication Technology

Comparative Analysis of the Performance of Feature Selection Methods in Diabetic Retinopathy Prediction Using Multilayer Perceptron Model

2025-01-08T12:05:20+00:00

Diabetic Retinopathy (DR), a leading cause of visual impairment among working-age adults, is increasing globally, necessitating effective predictive tools. Machine learning (ML) classifiers often struggle with high-dimensional datasets, making feature selection (FS) critical for improving predictive performance. This study evaluates the impact of FS techniques on the performance of an ML model for DR prediction using the MESSIDOR retinal dataset. Two FS methods, forward selection and variance threshold, were compared alongside a multilayer perceptron (MLP) classifier. The results showed that forward selection significantly enhanced MLP accuracy to 77.06%, outperforming the raw dataset (75.32%) and variance threshold (73.16%). The findings underscore the importance of appropriate FS in developing robust ML models. Integrating such models into clinical workflows could enhance early DR diagnosis, facilitate timely treatment, and reduce the risk of severe visual impairment, ultimately improving patient outcomes and healthcare efficiency.

Design and Implementation of a Computerized Tailoring Workshop Management System

2025-01-08T12:15:34+00:00

A Computerized Tailoring Workshop Management System is a software application designed to automate and streamline the operations of a tailoring workshop or a cloth production business. Despite the benefits, many tailoring workshops in Nigeria still lag in adopting computerized management systems due to limited resources and the complexity of the user interface. This study aims to design and implement a Computerized Tailoring Workshop Management System that can help address these challenges and that of the traditional manual management systems, to improve the overall efficiency and productivity of tailoring workshops in Nigeria. The Waterfall Model was employed in the system design and implementation. The implementation was done in phases, with testing and evaluation at each stage to ensure the system meets the requirements of a tailoring workshop. It was developed and implemented using Visual Basics programming language and Microsoft Access database. The system provides a user-friendly interface for tailors and customers to interact, track progress, and manage orders effectively and efficiently. The system's robust and scalable architecture will ensure reliability and performance, making it an ideal solution for tailoring workshops of all sizes.

Performance Analysis of selected Machine Learning Algorithms in the prediction of Man in the Middle in Internet of Things Environment

2025-01-08T12:25:03+00:00

Concerns have been expressed over Internet of Things (IoT) devices' growing prevalence and susceptibility to cyberattacks, namely Man-in-the-Middle (MitM) assaults. The performance of selected machine learning algorithms: Logistic Regression, Decision Trees, and K-Nearest Neighbors were analyzed and compared using accuracy, precision, recall, F1-score, and error rate using a dataset comprising normal and attacked data sets from Kaggle. According to the research findings, the Decision Tree algorithm outperformed other selected algorithms in terms of MitM attack prediction accuracy of 99.42% and a good balance between precision, F1-score, and recall, with the lowest error rate of 0. 0058. The results of the study improve the security and reliability of IoT applications by aiding in the creation of efficient MitM attack prediction systems for IoT environments. The findings also emphasize how crucial it is to choose the best machine-learning algorithm for a given IoT security task. Investigating the use of transfer other techniques in MitM attack detection for IoT contexts is one area of future research.

Data-Driven Framework for Adaptability Predictions Using Object-Oriented Metrics

2025-01-08T12:36:24+00:00

Software adaptability is a critical attribute in modern systems, enabling them to evolve with changing requirements and environments. Existing approaches to predicting adaptability often rely on subjective assessments or high-level architectural metrics, which may lack precision and scalability. This study aims to enhance adaptability prediction by integrating object-oriented metrics and machine learning models, addressing limitations of traditional methods. Decision Trees and Random Forests were employed to model relationships between object-oriented design metrics such as Coupling Between Objects, Lack of Cohesion in Methods, Depth of Inheritance Tree, and Number of Children and adaptability. A comparative analysis using Guava and Apache Commons Lang datasets revealed that the Random Forest model outperforms Decision Trees, achieving an F1-score of 0.9760 and a ROC AUC of 0.9858, highlighting its accuracy and feature importance. Key metrics like coupling and cohesion emerged as pivotal for adaptability prediction. This study contributes a robust, data-driven framework for adaptability prediction, offering valuable insights to developers for creating flexible and maintainable systems. These advancements improve software design practices, ensuring resilience and relevance of applications in dynamic technological landscapes.

Improved Blockchain Collaborative Integrity Verification Consensus with Consistent Nuanced Cues for Multicriteria Decision Making

2025-01-08T12:43:04+00:00

Decision making is an essential task that human undertake frequently. Decision is taken before actions. It may catalyze or preclude the success of an action depending on its rightness. Therefore, it is vital to the success of human endeavors. Naturally, making good decisions require personal wisdom or wisdom gotten from advice. Decision is often data driven because the facts on which the decision is based is usually generated from data. Genuine data generates correct facts. Computer is now being used to aid good decision. A computer assisted decision-making process is proposed in this work. The proposed method combined blockchain collaborative integrity verification consensus mechanism (CIVCM) and analytical hierarchy process (AHP) for thorough assessment of available alternatives to achieve a well-informed decision. This solution outperformed other solutions used for evaluating it by returning unambiguous results in the evaluation tests because of its working concept that avoids faults. The results shows that the solution is a better choice for high accuracy demanding applications.