SEMANTIC Features

SEMANTIC Features

1. represent basic conceptual component of meaning for any item i.e. word/phrase

2. full syntactic parsing or pragmatic analysis cannot yet be handled properly by NLP technology for unrestricted text

3. Semantic dependency

   • binary semantic
   • semantic modification relations
   • based on semantic dependency graph
   • semantic information when combined with lexical and syntactic information improved classification accuracy

4. Synonym and Hypernym of words

   • by applying latent semantic analysis to lexical features
   • to detect semantic similarities between words automatically

5. Functional features

   • inspired by Systematic Functional Grammar (SFG) theory
   • a set of functional features associated with words/phrases with semantic information
   • lexicon of words and phrases produced semi-automatically based on online thesauruses i.e. WordNet
   • each entry in the lexicon associated with a word/phrase with syntactic constraints and semantic properties
   • type of expansion
     • ELABORATION
     • EXTENSION
     • ENHANCEMENT
     • CONJUNCTION
     • CLARIFICATION
     • APPOSITION
   • measures
     • how many CONJUNCTION's expanded to ELABORATION's
     • how many ELABORATION's elaborated toCLARIFICATION's
   • E.g.: the word "specifically" used to identify CLARIFICATION