Automatic Sinhala Text Summarization for Government Gazettes using Abstractive and Extractive Methods

Abstract

During this new era, information is very accessible and the amount of text data from various sources has increased dramatically. However, most of them can distract the reader from the most important information due to using larger paragraphs, examples, complex arguments, grammar, and some vocabularies. Since time is one of the most important facts in the 21st century, people want to summarize these contexts and retrieve only the important information in a shorter time. The Gazettes are important to people in different way and there was no attempt on summarization solution for the area, this research emphases on the summarizing gazettes in the Sinhala language. This research solution is to provide summarized output for Sinhala gazettes by identifying the most important and relevant sentences based on linguistic and statistical features of a given text, using an abstractive and extractive approaches. Even though there are very few attempts done on the Sinhala Summarization this is the first attempt on summarizing Sinhala gazettes. The project was evaluated by machine summaries with the summaries created by the author. The system has been tested with 450 actual Sinhala gazettes and final results were attached in the Appendix section. Further, this provides a turning point for future researches on automatic text summarization in Sinhala language.