Extractive text summarisation using graph triangle counting approach: proposed method

Abstract

Currently, with a growing quantity of automated text data, the necessity for the con-struction of Summarisation systems turns out to be vital. Summarisation systems confine and condense the mainly vital ideas of the papers and assist the user to find and understand the foremost facts of the text quicker and easier from the dispensation of information. Compelling set of such systems are those that create summaries of ex-tracts. This type of summary, which is called Extractive Summarisation , is created by choosing large significant fragments of the text without making any amendment to the original. One methodology for generating this type of summary is consuming the graph theory. In graph theory there is one field called graph pruning / reduction, which means, to find the best representation of the main graph with a smaller number of nodes and edges. In this paper, a graph reduction technique called the triangle counting approach is presented to choose the most vital sentences of the text. The first phase is to represent a text as a graph, where nodes are the sentences and edges are the similarity between the sentences. The second phase is to construct the triangles, after that bit vector representation and the final phase is to retrieve the sentences based on the values of bit vector.