Semantic Representation Analysis (SRA) and Selected Applications

Xiangen Hu

Abstract

Semantic Representation Analysis (SRA) is a general framework for vector-based semantic analysis. Within this framework, semantics of natural language are represented in the form of Induced Semantic Structure (ISS). SRA has applications in information retrieval (IR), text analysis, and intelligent tutoring systems (ITS). In this talk, I will 1) introduce a mathematical model of SRA; 2) introduce and demonstrate a method that generates individualized domain-specific context sensitive semantic representation; 3) introduce and demonstrate learner’s characteristics curves (LCC) and its application in intelligent tutoring systems, 4) introduce and demonstrate Semantic Spectrum Analysis (SSA) and its applications in analyzing text streams in social media.