Method | Our proposal | FNCs | CMFMTS | MEMD |
|---|
Configuration | 20 | 40 | 60 | LSTM | ALSTM | MLSTM | MALSTM | C5.0 | RF | SVM | MEMD |
|---|
Accuracy | 0.95 | 0.98 | 1 | 0.86 | 0.82 | 0.71 | 0.78 | 0.89 | 0.9 | 0.79 | 0.73 |
- The row entitled “Method” shows the different methods to be compared: our proposal, the alternative based on deep learning (FNCs), another alternative based on feature extraction (CMFMTS), and the traditional method (MEMD). The row “Configuration” refers to the different configurations used for each method; in the case of our proposal, it refers to the number of variables selected (taking into account only the correlations), in FNCs the particular neural network used, in CMFMTS the classifier used, and MEMD has only been applied with one configuration
