A Study of Dictionary Based Korean Semantic Role Labeling

A semantic role is information used to clarify the role of entities in an event that a sentence describes, including agent, theme, experience, object, and location. Semantic role labeling (SRL) is a process that determines the semantic relation of a predicate and its arguments in a sentence and is an important factor in the semantic analysis of natural language processing, in addition to word sense disambiguation. To date, many manual semantic tagging tasks have been constructed; however, these tasks require a great deal of time and cost. To solve this problem, we propose a method for automatic SRL using frame files included in the Korean version of Proposition Bank (PropBank), which is one of the most widely used corpora. Frame files provide guidelines for PropBank annotators and include a list of framesets, which stand for a set of syntactic frames. First, we select the proper sense of the predicate from among multiple senses of the predicate in the frame files. Senses of the predicate are classified according to the semantic and syntactic properties of the predicate’s arguments. We collect the nouns in a sample sentence of a given sense; we also collect all of the nouns that appear in a given sentence. The semantic similarities between the nouns from the sample sentence and the given sentence are measured and the sense with the highest similarity value is selected. The frame information of the selected sense is used for SRL of the given predicate and its arguments.