The use of Semantic Role Labelling with Triangle-Graph Based Text Summarization

Text summarization aims to create a condensed version of a text, thus retaining the original ideas. In particular, the textual content on the web is increasing at an exponential pace. The ability to interpret valuable information from such a vast volume of data is an essential undertaking and requires an automated system to assist with the current information repository. Text summarization systems aim to help reducing content by retaining the relevant material and filtering out non-important sections of the text. There are two basic approaches in the text summarization systems. Single document summarization is outlined in the first method. In other words, the method uses one document as input and generates a summary version as output. An alternative approach is to take many documents as input and generate a single summary document as output. In terms of performance, the summarization systems are often classified into two groups. One method will be to remove exact sentences from the original document in order to create a summary production. An alternative would be a more nuanced approach in which the text rendered is rephrased version of the original document. A formal semantic representation is provided in this paper, which can capture the algorithms and the text meaning to allow mapping the text documents with the meaning representation. This semantic representation, which focuses on establishing relations amongst concepts of text constituents, is considered to enhance the performance of the single-document extractive summarisation processes. To see how text constituents (i.e. words, phrases and morphemes) are related, a graph semantic model (GSM) was used, which is built using a syntactic and semantic analysis. Semantic Role Labelling (SRL) is the fundamental relationship that a contributor has with the main verb in a sentence.