基於相依詞向量的剖析結果重估與排序(N-best Parse Rescoring Based on Dependency-Based Word Embeddings).