122
Voted
NAACL
2004
15 years 4 months ago
2004
We present a technique for augmenting annotated training data with hierarchical word clusters that are automatically derived from a large unannotated corpus. Cluster membership is...