Abstract:
In the paper, we discuss problem of semantic-syntactic parsing of texts in natural language. Review of approaches and methods for semantic-syntactic analysis is presented. We describe syntactic parser of Russian texts created using MaltParser and experimental research of features that influence performance of syntactic parsing of Russian texts. We describe system for semantic role labeling based on method of relational-situational analysis and present its evaluation results on Russian texts. We consider method for joint semantic-syntactic analysis and present its evaluation results on Russian texts. We also compare results of the system based on joint analysis and the system that performs separate consequent syntactic parsing and semantic role labeling.
Keywords:semantic-syntactic parsing, machine learning, syntactic parsing, semantic role labeling, parser.