Ensemble Learning
Mahnoor ijaz, Maham ehsan
what it is?
Ensemble learning is a
machine learning
technique that combines
multiple models to
produce a single, more
powerful predictive
model.
goal
The goal is to
improve accuracy,
robustness, and
generalization
compared to using a
single model.
benefits
Reduces errors like:
Bias: Fixes consistent mistakes by using different
models.
Variance: Makes predictions more stable and
reliable.
Overfitting: Avoids being too focused on training
data, making it better at handling new data.
key concepts
Aggregation:
Combines
Diversity: predictions •Base Models:
Models should
Decision Trees,
make varied
SVMs, etc.
errors
types
•Boosting (e.g.,
AdaBoost,
XGBoost):
•Bagging (e.g.,
Sequentially •Stacking: Uses
Random Forest):
corrects errors a meta-model
Reduces variance
to combine
by bootstrapping
predictions
advantages
• High accuracy and
robustness
• Reduces overfitting
• Versatile for various
datasets
disadvantages
• Computational cost
• Hard to interpret
• Overfitting risk if
poorly tuned
Thank you