AI Glossary Term
What does “Ensemble” mean?
Combining predictions from multiple models to improve robustness or accuracy.
How Ensemble works
Example
Three different models each predict whether a transaction is fraud. If two of three say yes, the system flags it. The combination is usually steadier than any single model because their mistakes are not identical.
Why it matters
Ensembles often win data-science competitions and improve reliability. They cost more to run, so teams weigh the gain against the extra compute.
Common questions
Why do ensembles work?
If models make different, partly independent errors, averaging or voting cancels some of them out.
What are common ensemble methods?
Bagging, such as random forests; boosting, such as gradient-boosted trees; and stacking, which trains a model to combine others.
Do ensembles always help?
Not if the models are too similar or if one is much worse than the rest. Diversity among members is key.