An Algorithmic Approach for Text Summarization

Abstract

With the advent of technology and increase in ease of access to digital devices, there has been a burgeoning increase in the flow and creation of information all across the internet. However, this has led to a lot of saturation and an increased amount of redundant information, which needs to be sorted through manually many times to find the important or pertinent information. But this ends up becoming a tiring and laborious task for many people. Hence, text summarizers, which are softwares that, when given a particular text input, summarize it effectively and output only the important parts in the text, while discarding the unimportant or redundant bits. A number of Text summarizers currently exist on the internet, for free and for a cost as well. Many of them are efficiently able to extract the important information and text out of the given input. However most of them are a single type of text summariser titled “Extractive Text Summarisers”, which, while fairly accurate, is based on a model that merely extracts the important texts from the given input as it is, without specifically phrasing it in any other way, or without semantic understanding of the text itself. This leads to some inaccuracies in the summarized texts, such as some irrelevant words being put in the summarized output, even though they are unimportant, or some important sentences or phrases missing out on being in the output, as they have not been discussed enough in the input. That is why, the proposed system of Query based Summarizer not only works on queries, but is also aimed to be an Abstractive Text Summarizer, which will use semantic and syntactic understanding of the given input to summarize it and deliver a clear, concise and accurate output.