Poverty Prediction Based on Economic Indicators 📊

The project aims to predict the probability of a country being poor using logistic regression, based on various economic indicators.

Dataset:

• Sourced from Kaggle

Features:

• MPI Urban: Multi-dimensional poverty index for urban areas.
• Headcount Ratio Urban: Percentage of the population listed as poor in urban areas.
• Intensity of Deprivation Urban: Average poverty depth for the poor in urban areas.
• MPI Rural: Multi-dimensional poverty index for rural areas.
• Headcount Ratio Rural: Percentage of the population listed as poor in rural areas.
• Intensity of Deprivation Rural: Average poverty depth for the poor in rural areas.

Target Variable:

• High Poverty Probability: The predicted probability of whether a country is likely to be below the poverty line. A higher probability indicates a higher likelihood of being in poverty.

Technical Overview:

• Input Data: Economic indicators for urban and rural areas.
• Model: Logistic regression to predict the probability of high poverty.
• Metrics: Accuracy, precision, recall, F1-score.

Highlights:

• Analyzes poverty across both rural and urban areas.
• Logistic regression model predicts the probability of high poverty.
• Decision boundary visualizations for urban and rural poverty indicators.

Key Steps:

1. Data Analyses.
2. Define target variable: Probability of high poverty based on economic indicators.
3. Train-test split (80% train, 20% test).
4. Model training using logistic regression.
5. Evaluate performance using accuracy, precision, recall, F1-score, and ROC-AUC.
6. Visualize decision boundaries for urban and rural indicators.

Explore the code to learn how machine learning aids in forecasting economic metrics! 🚀