A Reading Answering System Model for Vietnamese Language

Abstract

Based on our previous researches for building a Reading Answering System Model (RASM) which can read many simple news titles of ICTNEWS (http://ictnews.vn/) to answer related Vietnamese questions. The construction of RASM is based on an approach of computational semantics. In this paper we focus on introducing our approach to build the RASM, the general architecture, and functional operations of RASM. In particular, we present new experimental results to evaluate the performance of our system in practice. We tested the system on 8 datasets composing 403 Vietnamese testing questions, and a vocabulary of 1142 lexicons. In experiments, the precision of our system is 66.63%.