mscds

[mscds]

Machine Learning for Data Analysis-Week2-Running Random Forest:

Summary of Results

Model Accuracy and Performance: Accuracy: The Random Forest model achieved an accuracy of 94.34% on the test dataset, indicating a high level of correct predictions. Classification Report: For class 0 (e.g., life expectancy below the threshold), the model had a precision of 0.94, recall of 0.88, and an F1-score of 0.91. For class 1 (e.g., life expectancy above the threshold), the model demonstrated slightly better performance with a precision of 0.95, recall of 0.97, and an F1-score of 0.96. The weighted average for all metrics (precision, recall, F1-score) was 0.94, reflecting balanced performance across both classes.

Feature Importance: The most important features in predicting life expectancy (above or below the threshold) were: Survival to Age 65, Female (% of Cohort): This feature had the highest importance score (0.381), making it the strongest predictor in the model. Mortality Rate Under-5 (Per 1000): This was the second most important variable with an importance score of 0.201, highlighting its significant influence on life expectancy. Improved Sanitation Facilities (% of Population with Access): This feature also played a notable role, with an importance score of 0.139. Other important variables included fixed broadband subscriptions, access to electricity, and health expenditure per capita, though these had lower importance scores, indicating a more modest contribution to the model's predictions. The high accuracy and balanced performance metrics suggest that the Random Forest model is effective at predicting life expectancy based on the provided explanatory variables. The importance of "Survival to Age 65, Female" and "Mortality Rate Under-5" underscores the impact of health-related factors on life expectancy, consistent with expectations. Variables related to infrastructure (sanitation, broadband access) and economic factors (GDP per capita) also contributed, though to a lesser extent, indicating a multifaceted relationship between these predictors and life expectancy.