scSHARP Model¶
GCN Model¶
GCN model class propagates labels from component tool predictions. For more detailed model information see paper
- class scSHARP.sc_sharp.GCNModel(config_file, neighbors, target_types, seed=8, dropout=0.0)¶
class for label propagation GCN model
- construct_graph(X)¶
construct graph from data
- forward(X, edge_index=None)¶
forward pass through model
- predict(testloader)¶
predict with model
- to(device)¶
send layers to device
- train(trainloader, epochs, verbose=True, ignore_index_input=-1)¶
train model
- validation_metrics(testloader, train_nodes, test_nodes)¶
returns validation metrics (ROC, prC, accuracy, etc) test_nodes is array with indices of nodes whose labels were not used for training