- Open Notebook in Colab
This project predicts Life Expectancy using a WHO dataset (2000-2015). We analyzed how health indicators like HIV/AIDS & IMMUNIZATION, and socioeconomic factors like GDP influence longevity.
- Distributions: Used histograms and KDE plots to understand feature spreads.
- Correlations: A heatmap revealed that HIV/AIDS has a strong negative correlation with life expectancy.
- Missing Data: Handled via Simple Imputer.
Identified Schooling and Income Composition as the strongest positive predictors, while HIV/AIDS prevalence was the most significant negative influencer, providing data-backed evidence for public health prioritization.
We compared a baseline linear model against an advanced ensemble method.
| Model | MAE | R² Score |
|---|---|---|
| Linear Regression (Baseline) | 4.26% | 0.81 |
| Random Forest (Advanced) | 1.7% | 0.96 |
Conclusion: The Random Forest significantly outperformed Linear Regression, capturing non-linear relationships that the baseline missed.