Application Research of Ensemble Learning Frameworks
Kunkun Wang, Xianda Liu, Jianming Zhao, Hongwei Gao, Zhen Zhang
Abstract
In the field of artificial intelligence, machine learning is a hot research direction. With the improvement of computing power, machine learning has been better developed in the era of big data. A single machine learning algorithm has some shortcomings. Multiple simple base learning methods are combined to form an ensemble learning through a certain strategy, and meanwhile the performance of the ensemble learning will be greatly improved. The ensemble learning framework includes three integration modes: boosting, bagging, and stacking. This article introduces the development process of ensemble learning and studies the principles of the three ensemble methods, summarizes the existing ensemble learning algorithms, summarizes the characteristics and applications of the ensemble algorithms, and finally discusses the improvement and future development of the algorithms.