For more information see: https://www.kaggle.com/c/liverpool-ion-switching

The competition is basically about removing noise from a 1D signal.

IONCHANNEL_CONV1D_20200425.ipynb: Use of WAVENET or LADDERNET (U-NET) IONCHANNEL_LGB_20200502.ipynb: Use of lightgbm (Feature engineering + Boosting) IONCHANNEL_CATBOOST_20200502.ipynb: Use of CATBOOST (Feature engineering + Boosting)

Data: (folder) train.csv: original training data test.csv: original test data train_clean.csv: training data with drift removal. Obtained from Chris Deotte test_clean.csv: test data with drift removal. Obtained from Chris Deotte