Automatic Myanmar Text Summarization System

Abstract

Automatic text summarization is used as a tool to help people in reducing the time spent manually extracting the main ideas from text documents. If the natural disaster news is provided as the summary form including important and relevant information, people in management level can make comparisons and intelligent decisions quickly without exhausting energy by manually extracting the salient points. Moreover, for a normal user, automatic summary report of the disaster news makes them clear perception and fully awareness of the effects of the natural disaster by inspecting death toll and damage of the natural hazards. Therefore, this paper proposes Automatic Myanmar Text Summarization framework that is based on Information Extraction and practical implementation of this framework in summarizing natural disaster news which are in seven types: Earthquake, Flood, Landslide, Forest Fire, Tornado, Storm and Volcanic Eruption described in Myanmar Language. The two main components of the proposed framework, Myanmar Word Segmentation model based on Conditional Random Fields (CRFs) and Information Extraction Model using CRFs approach are also introduced.