Broad learning system based on ensemble learning
Yicheng Yan, Wei Guo, Wang Liwen · 2021
Aiming at the problem of poor stability of broad learning system (BLS), combining with the idea of ensemble learning, bagging BLS and stacking BLS algorithms are proposed. Firstly, the data set is sampled to get the data subset, and then the base classifier is trained by the broad learning system. Finally, the prediction results of the whole model are obtained by integrating the base classifiers. Through experiments on multiple data sets, the results show that the proposed algorithms achieve ideal results in image classification, and the two models are better than the single broad learning system in classification accuracy and variance, which shows the effectiveness of the proposed algorithm.