Automatic Text Summarization using
Wu-Palmer Measure and Graph based Sentence
Selection
Rishav Rakshit
1
, Chandralika Chakraborty
2
, Udit Kr. Chakraborty
3
and Bhairab Sarma
4
1
Microsoft Corporation, Hyderabad, India
Email: rishav.rakshit@microsoft.com
2-3
Sikkim Manipal Institute of Technology, (SMU), Majhitar, India
Email: {chandralika.c, udit.c }@smit.smu.edu.in
4
University of Science & Technology Meghalaya, Baridua, India
Email: sarmabhairab@gmail.com
Abstract—In this paper, an extraction based text summarization methodology is proposed that
aims to overcome issues faced with traditional extraction based summarizers using a
combination of testing semantic similarity using the Wu-Palmer Measure and a graph-based
sentence selection scheme for single document summarizations. The proposed approach is not
only independent of context, syntax and language but also promotes shorter summaries via its
sentence selection scheme. This method has been tested using the standard Reuters-21578
collection against widely used TextRank based summarizers and has shown promising results
for single-document summarizations.
Index Terms— Automatic, Text Summarization, Wu Palmer, Graph, extraction.
I. INTRODUCTION
The present decade is witnessing amelioration of connectivity. The massive reach of the internet and its rate of
growth has increased web-based content manifold. This has necessitated the development of text summarizers
which can be used to retrieve relevant information out of large texts. Text summarizers capable of delivering
succinct content, retaining the originality of the source document find large scale applications in the fields of
finance [1], journalism [2], medicine [3], internet search engines and even generic research [4].
A summary, according to Radev et.al [5] is defined as “a text that is produced from one or more texts, that
conveys important information in the original text(s), and that is no longer than half of the original text(s) and
usually, significantly less than that.” Automatic text summarization therefore can be taken as the process of
automatically producing the summary from a given text.
There exists two approaches to automatic text summarization, namely extractive summarization and abstractive
summarization. Extractive summaries consist of important sentences extracted from the parent text and copied
verbatim into the summary. Abstractive summaries, in contrast try to generate fresh sentences conveying the
most critical information from the text to be summarized. Traditional extractive summarizers do not guarantee a
coherent narration [6] but can represent an approximation of the content of the text with relative ease.
The current paper proposes a method of extractive summarization of single documents. The method proposed,
Grenze ID: 01.GIJET.8.1.521
© Grenze Scientific Society, 2022
Grenze International Journal of Engineering and Technology, Jan Issue