Inferring the Scope of Negation and Speculation Via Dependency Analysis

In the last few years negation and speculation scope classification systems have been developed successfully, most of them using machine–learning approaches. In this paper we present a competitive system that finds the scope of negation and speculation in sentences for English by means of syntactic dependency analysis without using machine–learning. A greedy algorithm first detects the cues (lexical markers that express negation and speculation), and second the words within the scope of these cues; finally, a Scope Finding algorithm uses these affected words to infer and to annotate the scope of the negation and the speculation. We tested our system with the Bioscope corpus, annotated with speculation and negation.