EmbeddingMachine

class EmbeddingMachine(seed=42, dimensions=128, max_iter=20)[source]

Bases: object

Tool to compute Personalized PageRank based embeddings.

Methods Summary

create_features(target, feature_definition)

Calculate the edge features based on node embeddings.

fit(pagerank_scores)

Train an embedding model.

Methods Documentation

create_features(target, feature_definition)[source]

Calculate the edge features based on node embeddings.

Parameters
  • target (DataFrame) – A dataframe of drug-drug interactions.

  • feature_definition (Callable[[ndarray, ndarray], ndarray]) – A Tigerlily edge feature computation function.

Return type

ndarray

Returns

Drug pair features for each edge.

fit(pagerank_scores)[source]

Train an embedding model.

Parameters

pagerank_scores (DataFrame) – A dataframe of the top-k personalized PageRank scores.

Return type

DataFrame

Returns

A node embedding for each source.