jola-ds/disease-diagnosis

This project predicts common illnesses in Nigeria (e.g., malaria, typhoid, pneumonia, gastroenteritis) using patient-reported symptoms and demographic data. # HealthLens: Disease Diagnosis Prediction Project DEMO ## Overview A machine learning project that predicts disease diagnosis from symptoms and patient data. This project demonstrates how data science can support early diagnosis and healthcare decision-making in low-resource settings. This project is a baseline solution with room for future optimisation, with current constraints and next steps detailed at the end. --- ## Problem Statement Nigeria faces a dual burden of communicable and non-communicable diseases. Malaria, tuberculosis, HIV/AIDS, and cholera remain widespread, while hypertension, diabetes, and cardiovascular disease are sharply rising (WHO, 2023; AHA Journals, 2023). Malaria alone contributes substantially to mortality and morbidity (Reuters, 2024). At the same time, access to timely medical intervention is limited outside major cities, worsening preventable illness and death (PMC, 2013; 2020). In low-resource settings, limited infrastructure, understaffing, and inequitable access to care remain persistent challenges, leading to gaps in diagnosis and treatment. Machine learning (ML) offers a way forward. Research has shown its potential in: - Predicting under-five mortality in Nigeria with up to 89.47% accuracy (BMC Med Inform Decis Mak, 2024) - Supporting HIV treatment adherence in Nigerian patients (PMC, 2024) - Enhancing outbreak prediction for diseases like cholera (Exploration Journals, 2023) Full Reference: `citations.md` ### Scope This project, therefore, addresses the urgent need for accessible, privacy-safe, and context-aware disease prediction tools. Using **domain-informed synthetic Nigerian hospital data**, a machine learning predictor was tailored to local healthcare realities. ⚠️ This tool is for educational purposes only. While not a substitute for clinical expertise, this system demonstrates how AI-driven models can serve as decision support tools for healthcare workers in low-resource settings—helping flag likely diseases …