-
Test-set Kappa for FFNN with FS (FFNN + FS) and without FS (FFNN + noFS).
SD: Standard Deviation. Trained with 60 epochs, learning rate of 0.1, one hidden fully-connected layer of size 100, and ELU activation. -
Test-set Kappa for RNN with optimization (RNN + FS + OPT) and without optimiz...
SD: Standard Deviation. -
Performance of the models, MAPE ratio of evaluation results for TAIEX futures.
Performance of the models, MAPE ratio of evaluation results for TAIEX futures. -
Performance of the models, MAPE ratio of evaluation results for SiMSCI futures.
Performance of the models, MAPE ratio of evaluation results for SiMSCI futures. -
r37980778c78--751f77e0a9bfc9966d7f3ae4cc9142b9
The motivation behind this research is to innovatively combine new methods like wavelet, principal component analysis (PCA), and artificial neural network (ANN) approaches to... -
Holm-adjusted <i>p</i>-values for the pairwise Wilcoxon post-hoc tests.
Non-significant differences (p > 0.05) in bold. -
r37980778c78--5941e8081fe72b02b593d7d06f28790c
Performance of the models, MAPE ratio of evaluation results for KOSPI futures. -
Test-set Kappa for FFNN with optimization (FFNN + FS + OPT) and without optim...
SD: Standard Deviation. -
r37980778c78--d021b843b76d1eb415fd7686d61e7c59
SD: Standard Deviation. Trained with 60 epochs, learning rate of 0.1, one hidden convolutional layer of 130 filters of size 5, and ReLU activation. -
r37980778c78--d6210ec597231984b560ff8fd14f744e
Return of the models, results for NIKKEI 225 futures. -
r37980778c78--aac8b49b7079c967e47a1529db630f02
Summary of parameter values for the GA used in CNN, FFNN, and RNN learning optimization. -
Return of the models, results for HANG SENG futures.
Return of the models, results for HANG SENG futures. -
Summary of parameter values for the GA used in the FS procedure.
Summary of parameter values for the GA used in the FS procedure.