StackedNode2Vec
|
Computes the Node2Vec representation of each node in a set of graphs. |
|
Find the Node2Vec representation of each node for each graph. |
|
|
|
- class graph_embeddings.algorithms.StackedNode2Vec(node_embeddings_size: int = 128, padding_value: int = 0, window: int = 5, walk_length: int = 100, n: int = 25, p: float = 1, q: float = 1, epochs: int = 10, algorithm: str = 'skip-gram', n_jobs: int = 1)[source]
Computes the Node2Vec representation of each node in a set of graphs.
- Parameters
- node_embeddings_size: int, default= 128
Size of the Node2Vec vector of each node
- padding_value: int, default=0
This number fills the rows in the graph’s matrix, corresponding to a node that is not present in an individual graph
- walk_length: int, default=100
Maximum length of each random walk
- window: int, default=5
Maximum distance between the current and predicted word within a sentence.
- n: int, default=25
Total number of random walks per root node
- p: float, default=1
Defines probability, 1/p, of returning to source node
- q: float, default=1
Defines probability, 1/q, for moving to a node away from the source node
- epochs: int, default=10
Number of iterations (epochs) over the corpus
- algorithm: str, {“skip-gram”, “CBOW”}
Training algorithm for Word2Vec.
- n_jobs: int, default=1
Use these many worker threads to train the model (=faster training with multicore machines).
- fit(graphs: List[Union[Graph, StellarGraph]])[source]
Find the Node2Vec representation of each node for each graph.
- Parameters
- graphs: List[Union[nx.classes.graph.Graph, StellarGraph]]
List of Networkx or StellarGraph objects
- Returns
- self: object
The whole StackedNode2Vec instance