Weight-aware random walks preserve edge weights in node embeddings better than unweighted or strength-based walks in most tested networks, with best synthetic correlations above 0.9 but real-world maximum 0.446.
We began by generating a random feature matrix, where each row represents a node and each column corresponds to a feature dimension
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SI 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Recovering link-weight structure in complex networks with weight-aware random walks
Weight-aware random walks preserve edge weights in node embeddings better than unweighted or strength-based walks in most tested networks, with best synthetic correlations above 0.9 but real-world maximum 0.446.