tech,18-3-P03-1002,bq of <term> features </term> ; and ( 2 ) <term> inductive decision tree learning </term> . The experimental results prove
other,8-4-P03-1002,bq results prove our claim that accurate <term> predicate-argument structures </term> enable high quality <term> IE </term>
other,9-1-P03-1002,bq we present a novel , customizable <term> IE paradigm </term> that takes advantage of <term> predicate-argument
other,9-2-P03-1002,bq new way of automatically identifying <term> predicate argument structures </term> , which is central to our <term> IE
other,15-1-P03-1002,bq paradigm </term> that takes advantage of <term> predicate-argument structures </term> . We also introduce a new way of
other,12-3-P03-1002,bq based on : ( 1 ) an extended set of <term> features </term> ; and ( 2 ) <term> inductive decision
other,18-2-P03-1002,bq structures </term> , which is central to our <term> IE paradigm </term> . It is based on : ( 1 ) an extended
tech,13-4-P03-1002,bq structures </term> enable high quality <term> IE </term> results . This paper proposes the
hide detail