A Novel Compression Algorithm for Hardware-Oriented Gradient Boosting Decision Tree Classification Model

Xiao Wang, Yafei Li, Yinqi Li, Kuizhi Mei

Abstract: Gradient boosting decision tree is a widely used machine learning algorithm. As the big data era comes, this algorithm has been applied to the multimedia fields. However, this algorithm suffers a lot when it comes to the mobility of multimedia application. In this paper, we propose a compression algorithm, GBDT Compression, to enhance the storage adaptability of a gradient boosting decision tree classification model. GBDT Compression introduces a new rule to pruning, addressing and encoding for a well trained gradient boosting decision tree model. By conduct GBDT Compression on small data sets, we show that this algorithm considerably reduces the space cost of original model up to dozens of times. Furthermore, as data increases and the original model grows, the compression rate also enhances. Which is meaningful for mobile multimedia application with limited memory.