Amalgam Based Multilingual Text Summarization For Devanagari Languages

The absence of efficient text summarizers for languages based on the Devanagari script like Hindi, Marathi and Sanskrit exhibit a conspicuous void in field of literature being devoid of the linguistic masterpieces. This study introduces a customized Multilingual Text Summarizer for Marathi and Hindi that makes use of a combination Python’s TextRank algorithm and Bag of Words (BOW) algorithm alongside extractive summarization techniques using a flask web application. By preserving the meaning and extracting most used words and sentences, our approach yields accurate summaries. The dataset used contains news articles and their summaries for testing our model. Empirical assessments using ROUGE approach confirms its efficiency in generating effective summaries with over 88% score.