Research and Application of AdaBoost Algorithm Based on SVM

Abstract

In this paper, we use AdaBoost algorithm as a framework to design a new classification algorithm. The algorithm combines the characteristics of SVM and AdaBoost algorithms. The SVM algorithm is used as a weak classifier algorithm. AdaBoost is used to cascade multiple weak classifiers to form a strong classifier, thus achieving more accurate classification. In the experiment, this paper uses SVM algorithm, AdaBoost algorithm and the algorithm of this paper to carry out several comparison experiments. The experimental results show that the algorithm can achieve higher accuracy classification in data sets of various sizes. At the same time, the algorithm is moderate in calculation speed, so the new algorithm proposed in this paper is better than the traditional SVM or AdaBoost classification algorithm.