A Hybrid Approach for Semantic Structure Annotating of Text

Facing the challenges of annotating naturally occurring text into semantic structured form for automatically information extracting, current Semantic Role Labeling (SRL) systems have been focusing on semantic predicateargument structure. Based on the Concept Description Language for Natural Language (CDL.nl) which aims to describe the concept structure of text by a set of pre-defined semantic relations, we develop a parser to add a new layer of semantic annotation of natural language sentences. The parsing task is a relation extraction process with two steps: relation detection and relation classification. We put forward a hybrid approach with different methods for two steps: firstly, based on dependency analysis, a rule-based method is presented to detect all entity pairs between each pair there exits a relationship; secondly, we use a featurebased method to assign CDL.nl relation to each detected entity pair with Support Vector Machine. We report our preliminary results on our manual dataset annotated with CDL.nl relations.