Automatic Semantic Role Assignment for a Tree Structure.

摘要: We present an automatic semantic roles labeling system for structured trees of Chinese sentences. It adopts dependency decision making and example-based approaches. The training data extracted examples are from the Sinica Treebank, which is a Treebank with role assigned each constituent. used 74 abstract including thematic roles, such as ‘agent’; ‘theme’, ‘instrument’, secondary ‘location’, ‘time’, ‘manner’ nominal modifiers. design assignment algorithm based on different features, head-argument/modifier, case makers, sentence structures etc. labels parsed Therefore practical performance depends good parser right achieves 92.71% accuracy in pre-structurebracketed texts considerably higher than simple method using probabilistic model head-modifier relations.