Statistics-based Semantic Role Labeling for Chinese menu corpus

Abstract

This paper describes our attempt at automatic Semantic Role Labeling for a Chinese specific corpus. We present a method to construct a domain specific corpus, a set of semantic roles that we will use, and their annotation. Additionally, we construct an ontology that we use in order to improve the accuracy of Chinese word segmentation and unknown word recognition. We propose two statistical methods for Semantic Role labeling and carry out experiments to compare them, and experimental results show that the methods achieve good performance.