Research on Sampling Diversity Method in Ensemble Learning Base on Margin
Gang Zhou, Fuliang Guo · 2019
From the deviation-variance decomposition and error-bifurcation decomposition, it is found that the diversity of base classifiers in ensemble learning is an important factor to improve the generalization ability and classification accuracy of ensemble learning. It can effectively improve the diversity of ensemble learning by the method of the minimum margin maximization in data samples. An optimal ensemble learning algorithm based on the minimum margin maximization was proposed for Based Bagging algorithm, which optimized the sampling method to maximize the minimum margin, and simplified the maximizing minimum margin problem to the base classifier weight adjustment problem. The minimum margin maximization of this type of data sample could implement by assigning higher weight to the base classifier of the correct classification error-prone data sample. Through data experiments, it is found that the optimization method has higher execution efficiency and generalization accuracy than the commonly used ensemble learning methods on most data sets.