1. Core Idea
- In ensemble learning, multiple models are trained and combined to improve the performance of the model.
- Weak and strong learners:
- Weak learner. A model that performs slightly better than random guessing.
- Strong learner. A model that performs well on a given task.
- Ensemble learning performs better when models are diverse and independent.
2. Voting
- Hard voting. The class with the most votes is predicted.
- Soft voting. The class with the highest average probability is predicted.
3. Popular Ensemble Algorithms