Machine Learning Techniques in Medical Diagnosis for Measles Symptoms in Jolo: Challenges and Opportunities
Main Article Content
Abstract
Despite being preventable, measles remains a major health threat in Jolo due to persistent challenges in diagnosis and healthcare access, especially in resourcelimited settings. This study assessed these challenges—specifically in diagnosis and medical data accessibility, healthcare workforce and communication, and infrastructure and ethical concerns—while designing a machine learning (ML)- based application tailored to local needs. Using a mixed-methods research design, the study combined quantitative and qualitative data to evaluate the system’s feasibility and effectiveness. Developed through Agile methodology, the application underwent iterative refinement to remain responsive to user needs. Quantitative data, gathered through system testing and Likert-scale surveys among IT experts and healthcare workers, showed a “very acceptable” level of user satisfaction across functionality, usability, reliability, and performance. The ML model achieved approximately 93% validation accuracy over 50 training epochs, indicating strong generalization. Qualitative insights from interviews and observations revealed the system’s cultural relevance, usability, and addressed ethical concerns. The results highlight the application’s potential to improve diagnostic accuracy, enable early detection, and expand access to care in underserved areas. Although reliability and performance had modest areas for improvement, stakeholders strongly supported the tool’s integration into local healthcare systems. Overall, the study demonstrates the viability and promise of ML-based solutions in addressing measles and broader healthcare challenges in remote communities.