Open GouhaoC opened 2 weeks ago
In the experiments regarding the number of base classifiers, Tables 5 and 6 show that when (C=1), the GEN model performs best on the training set, validation set, and test set. What is the purpose of ensembling then?
Thank you for your comment. Tables 5 and 6 are mainly to verify the overfitting of GEN, and the complexity, where a lot of C is omitted, leading to the fact that it looks like GEN performs best when C=1. In fact, Table 2 is the optimal hyperparameters provided in this paper, please refer to Table 2 when you need to reproduce the related experiments.
Thank you for your answer. There are some problems when I run the code. Is there anything that needs special modification?
In the experiments regarding the number of base classifiers, Tables 5 and 6 show that when (C=1), the GEN model performs best on the training set, validation set, and test set. What is the purpose of ensembling then?