Life Expectancy Prediction Project

Overview

This project aims to predict life expectancy using socio-economic and health-related factors, leveraging WHO data from 2000 to 2015. The dataset consists of approximately 3000 data points with 22 features.

Data Preprocessing

Key steps in data preprocessing included:

• Handling null values using a Simple Imputer (median).
• Label encoding categorical variables.
• Splitting the dataset in an 80:20 train-test ratio.
• Normalizing the data.
• Addressing multicollinearity by removing highly correlated features.

Feature Selection

Feature selection was performed using wrapper methods like forward selection and backward elimination across various models:

• Linear Regression
• Decision Tree
• Random Forest
• Xgboost

Model Selection and Validation

• Hyperparameter tuning was conducted using GridSearchCV.
• Best performing models: Random Forest and XGBoost with forward selection, achieving an R-squared of ~0.96.
• 

Conclusion

The project demonstrates the effectiveness of machine learning in predicting life expectancy, with significant implications for healthcare policy and resource planning.