Skip to content

ALEX95GOGO/CAGNN

Repository files navigation

CAGNN

A graph neural network that can generate connectivity via end-to-end training

Training

python train.py --save_dir ./save/ --max_seq_len 10 --do_train --num_epochs 30 --metric_name F1 --use_fft --lr_init 1e-3 --num_rnn_layers 2 --rnn_units 32 --max_diffusion_step 2 --num_classes 2 --graph_type combined --num_nodes 30 --input_dim 3 --data_augment --test_batch_size 320 --train_batch_size 200 --eval_every 1--rand_seed 0

Set the num_node to the number of channels in EEG, input_dim equal to the length of feature vector, num_classes equal to the number of classes in the input

Visualization

python plot_explanability.py

This will plot the explanable figures highlighting the sequence most relevant to the network prediction

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages