Another ensemble technique that sequentially trains models, each one focusing on the errors of the previous ones. Boosting combines multiple “weak” learners into a strong composite model by weighting them according to performance. Like bagging, boosting (e.g. AdaBoost, gradient boosting) can improve accuracy of price predictions from noisy forex data. Both bagging and boosting exemplify how combining several models often yields more robust signals for forex trading.

Back to Robots Glossary Index