| 1 |
Author(s):
Kazeem Sodiq, Otapo Akeem, Lateef Saliu, Peter Balogun, Latifat Oderinu, Solomon Salau, Oyesola Awoyoola.
Page No : 1-19
|
Review on Prediction of Heart Attack Using Traditional and Artificial Intelligence Approaches.
Abstract
Heart attack, clinically referred to as acute myocardial infarction (AMI), remains one of the leading causes of global mortality, largely due to delayed diagnosis and the complex interplay of cardiovascular risk factors. Early and accurate prediction of heart attacks is critical for reducing morbidity, guiding preventive interventions, and improving patient outcomes. Traditional diagnostic and predictive approaches—such as risk-score models, biomarker analysis, and imaging techniques—provide valuable insights but often lack the precision, adaptability, and predictive strength required for real-time clinical decision-making. Recent advancements in artificial intelligence (AI), particularly deep learning, have demonstrated remarkable potential in identifying subtle, nonlinear patterns within large medical datasets that may be overlooked by conventional methods.This paper presents a comprehensive and systematic review of traditional, machine learning, and deep learning techniques used for heart attack prediction. Key algorithms—including Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM), Deep Belief Networks (DBN), and Denoising Autoencoders (DAE)—are examined with respect to model design, dataset characteristics, evaluation metrics, interpretability, and clinical readiness. Findings from previous studies indicate that hybrid deep learning models consistently outperform standalone methods, achieving higher accuracy, sensitivity, and robustness. Despite these advancements, significant gaps remain in explainability, dataset diversity, real-world validation, and integration into clinical workflows.The review highlights the potential of AI-driven predictive systems to support early detection, reduce healthcare costs, and promote personalized cardiovascular care. It concludes by outlining future research directions focused on interpretable hybrid models, multimodal datasets, and deployment-ready systems capable of enabling proactive, patient-centered heart attack prevention.
| 2 |
Author(s):
Olawunmi Asake Adebanjo, Adebowale Oluwasegun, Anuriam Isaac, Adewuyi Joseph Oluwaseyi, Agoha Emmanuel.
Page No : 20-34
|
Design and Implementation of an AI-Powered Task Management System (TaskWise).
Abstract
Managing tasks across academic, personal, and professional life remains a persistent struggle for most people, even with the growing number of digital tools available. The problem is not a shortage of apps—it is that the vast majority of them still require users to do most of the work manually: typing out tasks, setting deadlines one at a time, and remembering to check back. There is little to no intelligence built into these platforms to anticipate what a user might need next. This paper introduces TaskWise, a task management application that takes a different approach. Rather than forcing users into rigid input forms, TaskWise lets them type or speak commands in plain English—for example, “Remind me to submit my assignment by Friday”—and the system interprets those instructions to create, schedule, and organise tasks on their behalf. The application was built on a three-tier architecture: React and Tailwind CSS handle the frontend interface, Firebase Cloud Functions drive the backend logic, and Firestore provides real-time data synchronisation across devices. We also incorporated an AI-powered conversational assistant that can generate study plans, suggest tasks based on context, and respond to user queries in a natural, conversational tone. Testing with a small group of university students showed that users could create tasks significantly faster with the natural-language input compared to traditional form-based entry, and the System Usability Scale score came in at 82 out of 100. While the current version is still a prototype with clear limitations—particularly around offline support and team collaboration—TaskWise demonstrates that combining AI with cloud infrastructure can meaningfully reduce the friction involved in day-to-day task management.
| 3 |
Author(s):
Adegbola Adesoji, Umaru Victor, Odebunmi Adeoluwasubomi, Olaitan Temitayo.
Page No : 35-49
|
MedLing: A Multilingual Natural Language Processing Framework for Medical Diagnosis Support System.
Abstract
Access to digital healthcare in Nigeria remains limited by language barriers, as most medical platforms are designed primarily in English, while a large portion of the population communicates in indigenous languages. This gap often prevents users from accurately describing symptoms, restricting effective communication and access to appropriate healthcare services.
This study presents a multilingual medical diagnosis system capable of accepting symptom descriptions in English, Yoruba, Igbo, Hausa, and Nigerian Pidgin. The system generates ranked differential diagnoses with clinically grounded descriptions and treatment recommendations delivered in the user’s preferred language. Using advances in Natural Language Processing, the framework employs the Aya Expanse 8B model, deployed through Ollama, to translate indigenous inputs into English for downstream analysis.
Treatment recommendations are generated through a Retrieval-Augmented Generation (RAG) architecture integrating WHO Essential Medicines guidelines, Federal Ministry of Health (FMOH) protocols, NIH MedlinePlus, and OpenFDA resources. Output reliability is improved through deterministic description generation, a three-tier language refinement pipeline, and hallucination detection algorithms, while Supabase manages authentication and interaction history.
Evaluation results show strong performance, achieving 89% Top-3 accuracy, low hallucination rates, and high linguistic reliability, demonstrating the effectiveness of the proposed hybrid deterministic-RAG framework for multilingual healthcare support in low-resource African languages.
| 4 |
Author(s):
Adesoji Adedeji Adegbola, Adebayo Olamide Solomon, Gafaru King Oluwashina, Adewale Saheed Eniola.
Page No : 50-67
|
Development of a Multi-Channel Emergency Response System.
Abstract
Across many parts of the world, the escalating frequency of insurgency, banditry, and kidnapping has stretched emergency response infrastructure to its limits, bringing to light deep-seated inefficiencies in how security threats are communicated and managed. Traditional emergency response systems frequently fail to preserve coordinated and timely communication, producing outcomes that are operationally inadequate due to their centralized, manual, and reactive processing model.
In this study, we presents the development of a mobile-based, multi-channel emergency security alert system built on a real-time location tracking and automated agency notification pipeline. The system integrates React Native for cross-platform mobile interaction, Django REST Framework for backend logic and API communication, and a multi-channel delivery model transmitting alerts via push notifications, SMS, and email. A custom acknowledgment feedback mechanism manages response status updates, while a layered architecture maintains separation of concerns across components enabling scalability and modular extensibility. The system is evaluated using response time, alert delivery reliability, and usability metrics.
Results confirm that the system fulfills its core functional requirements reliably. Unit and system testing returned a full pass rate across all critical test cases, encompassing authentication, priority computation, agency assignment, and end-to-end alert delivery. Multi-channel notification was successfully achieved across push notifications, SMS, and email, with alert status tracking progressing accurately through all defined response stages. These findings demonstrate the viability of a mobile-first, automated approach to real-time security alert management in high-risk environments.
| 5 |
Author(s):
Eze Monday, Ebiesuwa Oluwaseun, Oyebola Akande, Okesola Kikelomo I., Ojo Abosede Ibironke, Mgbeahuruike Emmanuel O.
Page No : 68-86
|
From Black Box to Clinical Trust: A Conceptual Review of Explainable and Lightweight Deep Learning for Spine Disease Detection and Segmentation.
Abstract
Deep learning models have made significant contributions to the accurate detection of spinal anomalies. However, persistent challenges related to model transparency, explainability and computational demands continue to hinder the real-time deployment of AI-driven solutions in clinical settings.
This review examines contributions from researchers on improving the explainability and transparency of deep learning models for spine image analysis, with the aim of identifying promising approaches and informing future research directions. Specifically, it explores the potential synergy between explainable AI techniques and lightweight models, with the expectation that such integration will yield models that are simultaneously accurate, interpretable, and clinically deployable.
Concepts covered include deep learning architectures applied to spine imaging tasks such as classification and segmentation; lightweight model design strategies; and categories of explainability techniques including Grad-CAM, LIME, and attention mechanisms. Constraints to the full clinical adoption of AI solutions in spine imaging are also discussed.
Key findings highlight several gaps in the research field. Dataset limitation issue, absence of standardised metrics for evaluating model interpretability, and challenges surrounding clinical acceptability. These gaps point to research direction that calls for greater collaboration between healthcare professionals and AI researchers, to develop spine imaging solutions that are both explainable and practically usable in clinical environments.
