🏥 ML Prediction of Acute Respiratory Infections (ARI) in Bangladesh

Rural vs. Urban Analysis for Under-Five Children

📌 Overview

This project applies machine learning (ML) techniques to predict Acute Respiratory Infections (ARI) among children under five in Bangladesh, comparing urban and rural areas. ARI remains a leading cause of child mortality, and early detection can help policymakers and healthcare providers take preventive measures.

---

📂 Dataset

• Source: ARI dataset from Bangladesh
• Format: CSV
• Key Features:
  • Demographic variables (age, gender, household location)
  • Socioeconomic factors (parental education, income levels)
  • Environmental factors (air quality, household cooking fuel)
  • Health indicators (previous infections, vaccination status)
  • Target Variable: ARI (0 = No ARI, 1 = ARI Present)

---

🏗️ Methodology

🔍 Data Preprocessing

• Handling missing values
• Encoding categorical variables
• Feature scaling using StandardScaler
• Splitting data into training & testing sets

📊 Exploratory Data Analysis (EDA)

• Class distribution analysis
• Correlation heatmaps
• Feature importance analysis
• Urban vs. Rural ARI prevalence comparison

🔧 Machine Learning Models

The following models were implemented and evaluated:

Model	Description
Logistic Regression	Baseline model for classification
Random Forest	Handles non-linearity and feature importance
Gradient Boosting (XGBoost)	Advanced tree-based method for better accuracy
Neural Networks	Deep learning approach for complex feature interactions

🏆 Model Evaluation

• Accuracy, Precision, Recall, F1-Score
• ROC-AUC Curve for performance comparison
• Feature importance ranking

---

🚀 Installation & Usage

🔥 Prerequisites

Ensure you have the following installed:

• Python 3.8+
• Pandas, NumPy, Matplotlib, Seaborn
• Scikit-learn, XGBoost, TensorFlow

📥 Clone Repository

git clone https://github.com/your-username/ML-Prediction-of-Acute-Respiratory-Infections.git
cd ML-Prediction-of-Acute-Respiratory-Infections