# Feature Learning Viewpoint of AdaBoost and a New Algorithm

## TL;DR

This paper offers a feature learning perspective on AdaBoost's resistance to overfitting and introduces a new AdaBoost+SVM algorithm that explains this phenomenon through feature space analysis.

## Contribution

It proposes a novel AdaBoost+SVM algorithm and provides a theoretical explanation for AdaBoost's overfitting resistance from a feature learning viewpoint.

## Key findings

- AdaBoost+SVM effectively explains AdaBoost's resistance to overfitting.
- Increasing feature dimension does not degrade SVM performance, supporting the theory.
- The new algorithm offers a more understandable framework for AdaBoost's behavior.

## Abstract

The AdaBoost algorithm has the superiority of resisting overfitting. Understanding the mysteries of this phenomena is a very fascinating fundamental theoretical problem. Many studies are devoted to explaining it from statistical view and margin theory. In this paper, we illustrate it from feature learning viewpoint, and propose the AdaBoost+SVM algorithm, which can explain the resistant to overfitting of AdaBoost directly and easily to understand. Firstly, we adopt the AdaBoost algorithm to learn the base classifiers. Then, instead of directly weighted combination the base classifiers, we regard them as features and input them to SVM classifier. With this, the new coefficient and bias can be obtained, which can be used to construct the final classifier. We explain the rationality of this and illustrate the theorem that when the dimension of these features increases, the performance of SVM would not be worse, which can explain the resistant to overfitting of AdaBoost.