Research Article
A Novel Algorithm of Machine Learning: Fractional Gradient Boosting Decision Tree
@INPROCEEDINGS{10.1007/978-3-031-18123-8_58, author={Kangkai Gao and Yong Wang}, title={A Novel Algorithm of Machine Learning: Fractional Gradient Boosting Decision Tree}, proceedings={Multimedia Technology and Enhanced Learning. 4th EAI International Conference, ICMTEL 2022, Virtual Event, April 15-16, 2022, Proceedings}, proceedings_a={ICMTEL}, year={2022}, month={10}, keywords={Fractional calculus Gradient descent method Ensemble method GBDT}, doi={10.1007/978-3-031-18123-8_58} }- Kangkai Gao
Yong Wang
Year: 2022
A Novel Algorithm of Machine Learning: Fractional Gradient Boosting Decision Tree
ICMTEL
Springer
DOI: 10.1007/978-3-031-18123-8_58
Abstract
The gradient boosting decision tree is a commonly used and effective ensemble machine learning method. In this paper, a factional gradient boosting decision tree scheme is first proposed with several loss functions discussed. This scheme implies that different algorithms can be established when adopting different fractional order gradient methods. It is then shown that with satisfactory convergence of fractional gradient descent algorithms, the proposed algorithms can train models with required accuracy in less number of iterations. Finally, several examples are provided to demonstrate the accuracy and efficiency of the algorithms.
Copyright © 2022–2025 ICST
