This is the implementation code for the paper "Trainable Undersampling for Class-Imbalance Learning" published in AAAI2019
load dataset for given task; define some evaluation functions.
implement the policy; perform model training.
similar to trainer.py, but trains the policy on gradually increasing data set.
generate synthetic data; choose supervised classifier and its corresponding hyper-parameters; get results reported in Table 1 of the paper.
choose supervised classifier and its corresponding hyper-parameters; plot classification boundaries on original dataset and the sampled dataset with our proposed method as reported in Figure 1 of the paper.
choose supervised classifier and its corresponding hyper-parameters on page dataset; apply typical data sampling methods to this dataset and the chosen classifier.
similar to page.ipynb but performs on the spam message dataset.
similar to page.ipynb but performs on the vehicle dataset.
similar to page.ipynb but performs on the creditcard dataset.