Paper Review: Random Forests

A guided review of Breiman’s random forest paper, ensemble strength, correlation, robustness, and variable importance.

Summary

Breiman described random forests as a combination of tree predictors built with randomization. The paper connects generalization error with tree strength, tree correlation, noise robustness, and variable importance.

Strengths

• It links ensemble performance with strength and correlation.
• It provides internal estimates for error and variable importance.
• It covers classification and regression use cases.

Limitations

• A forest is less interpretable than one small decision tree.
• Tree count and feature sampling still need practical tuning.
• Variable importance can be misleading when features are correlated.

Conclusion

Random forest remains a durable tabular baseline because it combines strong performance, robustness, and useful diagnostic tools.

Reading guide

Read the abstract first, then focus on the strength-correlation discussion and the variable importance section.