# Boosting

## What is Boosting?

Machine learning algorithm to create classification model

### No Algorithm is Perfect

No single algorithm is going to be most accurate. Every algorithm has its weaknesses and draw backs. What if we could combine all the results of several classifiers?

Have several basic classifiers (called v classifiers or learners) and combine them to form a strong classifier.

Decision Stump: A simple decision that uses only one feature. Basically a horizontal or vertical line in your data set (not diagonal, that would be multiple dimensions).

$h(x;\theta_j)=sign(w_jx_j+b_j)$

Boosting is how to combine all these week learners.