Data Process#
- fedgraph.data_process.load_data(dataset_str: str) tuple[source]#
Loads input data from ‘gcn/data’ directory and processes these datasets into a format suitable for training GCN and similar models.
- Parameters:
dataset_str (Name of the dataset to be loaded.) –
- Returns:
features (torch.Tensor) – Node feature matrix as a float tensor.
adj (torch.Tensor or torch_sparse.tensor.SparseTensor) – Adjacency matrix of the graph.
labels (torch.Tensor) – Labels of the nodes.
idx_train (torch.LongTensor) – Indices of training nodes.
idx_val (torch.LongTensor) – Indices of validation nodes.
idx_test (torch.LongTensor) – Indices of test nodes.
Note
ind.dataset_str.x => the feature vectors of the training instances as scipy.sparse.csr.csr_matrix object; ind.dataset_str.tx => the feature vectors of the test instances as scipy.sparse.csr.csr_matrix object; ind.dataset_str.allx => the feature vectors of both labeled and unlabeled training instances (a superset of ind.dataset_str.x) as scipy.sparse.csr.csr_matrix object; ind.dataset_str.y => the one-hot labels of the labeled training instances as numpy.ndarray object; ind.dataset_str.ty => the one-hot labels of the test instances as numpy.ndarray object; ind.dataset_str.ally => the labels for instances in ind.dataset_str.allx as numpy.ndarray object; ind.dataset_str.graph => a dict in the format {index: [index_of_neighbor_nodes]} as collections.defaultdict object; ind.dataset_str.test.index => the indices of test instances in graph, for the inductive setting as list object.
All objects above must be saved using python pickle module.
- fedgraph.data_process.normalize(mx: csc_matrix) csr_matrix[source]#
This function is to row-normalize sparse matrix for efficient computation of the graph
- Parameters:
mx (sparse matrix) – Input sparse matrix to row-normalize.
- Returns:
mx – Row-normalized sparse matrix.
- Return type:
sparse matrix
Note
Row-normalizing is usually done in graph algorithms to enable equal node contributions regardless of the node’s degree and to stabilize, ease numerical computations.