I have added some new methods eg. LightGBM, Separate decomposition methods...
(1) You can try SVM and LightGBM method by cl.sklearn_perdictor().
(2) More separate decomposition methods like SVMD, SEMD, SCEEMDAN are available.
(3) Have fixed some minor problems.
Feel free to contact me for more feedback! 欢迎提问和反馈报错~
CEEMDAN_LSTM
(1) Modify the cl.sklearn_perdictor() of sr.svm_sklearn_perdict().
(2) Modify the cl.load_dataset().
(3) Modify the rolling method.
(4) Modify the VMD decomposition.
(5) Add the OVMD decomposition method by optuna.
(6) Add the SVMD.
(7) Delete the CuDNNLSTM and CuDNNGRU.
(8) Add the out-of-sample forecasting for the rolling method.