search

dalinzhangzdl / BCI_MI_Wavelet_CNN

Using wavelet transform to extract time-frequency features of motor imagery EEG signals, and classify it by convolutional neural network
MIT License
103 stars 28 forks source link

Model not learning for BCI Competition 4 2a dataset #6

Open priyanshu7401 opened 1 year ago

priyanshu7401 commented 1 year ago

Hi there, i tried your model for BCI 4 2a dataset and for me, it is not learning as the maximum accuracy that was achieved was 0.34. Can you please provide the notebook for the BCI competition 4 2a dataset?

Anchaliya75 commented 1 year ago

@priyanshu7401 can you share your work mine accuracy is also around 34%I tried 2d cnn Training data accuracy is 99% but testing data is around 33%