A Naive Extractive Text Summarizer for Assamese Language

Text summarization is a pertinent problem of Natural Language Processing that deals with finding the summary of an input text or a set of text by using extractive or abstractive technique to find the important points of the text while maintaining the context and coherence. This paper describes the use of a naive extractive technique to summarize texts written in Assamese language, which is an Indo-Aryan language. The process starts with preprocessing of text, which is done manually and the term frequency of the words are considered to find the sentence scores, which are used to filter out the top-ranked sentences, based on which the summary is prepared. The feature of the system is that a user can insert their own document to summarize as well as can select the number of sentences in which they will get the summary. A set of 12 text is used to experiment on the code and although it does not incorporate semantic information into account, the results are quite promising, showing the validity of this naive extractive summarization approach.