ARIIK-ANTHONY/IndabaX-South-Sudan-2025-AI-Hackathon

# IndabaX South Sudan 2025 AI Hackathon - Blood Disease Classification ## Team: CodeNomads **Member**: ARIIK ANTHONY MATHIANG **Member**: JONGKUCH CHOL ANYAR **Member**: JOK JOHN MAKER ### Achievement **Target Accuracy**: 0.98559+ --- ## Project Overview Advanced machine learning solution for blood disease classification using ensemble methods and medical domain expertise. This project tackles the critical healthcare challenge of automated blood disease diagnosis using state-of-the-art AI techniques. ### Problem Statement Develop an AI model to classify blood diseases based on medical laboratory test results, enabling faster and more accurate diagnosis in healthcare settings. --- ## Key Features - Medical Domain-Specific Feature Engineering: Custom health indicators and blood ratios - Ensemble ML Models: Random Forest + Gradient Boosting with soft voting - Strategic Prediction Distribution Matching: Optimized for competition scoring - Championship-Level Performance: Targeting 98.5%+ accuracy --- ## Dataset Information ### Classes Distribution - **Diabetes** (60.5%) - 294 samples - **Anemia** (17.3%) - 84 samples - **Thalassemia** (9.9%) - 48 samples - **Heart Disease** (8.0%) - 39 samples - **Thrombocytopenia** (3.3%) - 16 samples - **Healthy** (1.0%) - 5 samples ### Features (24 Medical Parameters) - **Blood Composition**: Glucose, Cholesterol, Hemoglobin, Platelets - **Blood Cells**: White/Red Blood Cells, Hematocrit, MCV, MCH, MCHC - **Metabolic**: Insulin, BMI, HbA1c, Triglycerides - **Cardiovascular**: Blood Pressure, Heart Rate, Troponin - **Liver Function**: ALT, AST - **Kidney Function**: Creatinine - **Inflammation**: C-reactive Protein - **Cholesterol Profile**: LDL, HDL --- ## Project Structure ``` IndabaX-South-Sudan-2025-AI-Hackathon/ ├── client/ # Frontend React application │ ├── src/ # Source code │ │ ├── components/ # Reusable UI components │ │ ├── pages/ # Appl …