Utilizing Neural Networks for Early Prediction of Pneumococcal Disease: A Case Study in Bonny Island, Nigeria

Abstract

Pneumococcal disease, caused by Streptococcus pneumoniae, poses a significant health challenge, particularly in resource-limited settings like Bonny Island, Nigeria. This study employs neural networks and artificial intelligence to predict pneumococcal disease, addressing the critical need for early diagnosis and intervention. Methodologically, the research encompasses data collection, cleaning, correlation analysis, and model development, ensuring a robust system for early disease prediction. By analyzing demographic, clinical, and environmental factors, the study identifies significant predictors of pneumococcal disease risk. In comparison with Random Forest and Support Vector Machines trained on the same data, the neural network achieved 100 percent accuracy, recall, precision, and f1 scores. The integration of the neural network model into a web application facilitates real-time predictions, enabling healthcare providers to input symptoms and receive immediate diagnostic insights. This approach enhances timely interventions, potentially reducing morbidity and mortality associated with pneumococcal disease. Despite challenges like data quality and integration, the findings demonstrate the efficacy of AI-driven models in improving public health outcomes. The deployment of such models in Bonny Island underscores their practicality and scalability, paving the way for broader applications in similar contexts. Ultimately, this study not only advances understanding of pneumococcal disease epidemiology in Bonny Island but also contributes to global efforts in enhancing healthcare delivery through innovative technological solutions. Future research should focus on continuous model refinement and validation with larger datasets to further improve accuracy and reliability.