A Survey on Automatic Text Summarization

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 09 Issue: 03 | Mar 2022 www.irjet.net p-ISSN: 2395-0072
© 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 846
A Survey on Automatic Text Summarization
Vandit Mehta1, Khushi Patel2, Arham Shah3, Sejal Thakkar4
1-3Department of Computer Engineering, Indus University,Ahmedabad, India.
4Assistant Professor CE Dept. IITE Indus University
---------------------------------------------------------------------***----------------------------------------------------------------------
Abstract - Text Summarization is a Natural Language
Processing (NLP) method that extracts and collects data
fromthe source and summarizes it. Text summarization has
become a requirement for many applications since
manually summarizing vast amounts of information is
difficult, especially with the expanding magnitude of data.
Financial research, search engine optimization, media
monitoring, question-answering bots, and document
analysis all benefit from text summarization. This paper
extensively addresses several summarizing strategies
depending on intent, volume ofdata, and outcome. Our aim
is to evaluate and convey an abstract viewpoint of the
present scenario research work fortext summarization.
Keywords: Natural Language Processing, Text
Summarization, Abstractive Summary, Extractive
Summary.
1.INTRODUCTION
Since the advancement in the utilization of the Internet
hasexpanded, tremendous volumes of data are generated.
Most of the generated data is unstructured, so manually
extracting meaningful data from it is challenging [1].
Humans have a constrained ability to comprehend and
extract useful information from large amounts of data. It
takes a long time for them to grasp the essence of the
content. As a result, automatic summarization is a well-
known wayofaddressingsuch challenges [2].
The objective of text summarization is to gather
prominent information from the source by filtering and
providing a succinct summary [1]. To date, several
techniques for text summarization have been developed.
Text summarization techniques can be broadly classified
into four categories: input, output, content and purpose.
There are single and multi-document summary options
based on the number of documents. Meanwhile, the
extractive and abstractive outcomes are based on the
summary results. In contrast, generic and query-based
depend on the purpose [3]. On the other hand, it is
divided into indicative and informative based on the
content.
The internet is abundant with raw text from several
sources,and genres are typically unstructured, noisy, and
unsuitable for summary processing [4]. Text pre-
processing refers to the process of cleaning and
standardizing the unstructured data. It is a necessary
step before we can begin text summarizing. The five
components of text pre-processing are tokenization, lower
casing, stop words removal, stemming, and lemmatization.
Fig -1: Steps of Text Pre-processing
2.TYPES OF TEXT SUMMARIZATION
I. Extractive vs. Abstractive
Extractive text summarization works by selecting important
words, phrase or sentences, and concatenating them to
form a meaningful summary. Sentences are chosen based
on statistical and linguistic characteristics [5]. Whereas
abstractive summarization uses linguistics to examine
and interpret the text, and then constructs new sentences
and words while maintaining the source's content in a
comprehensible summary [6].
II. Single Document vs. Multi Document
Single document summarization (SDS) accepts a single
document, while multiple document summarization
(MDS) accepts several documents as an input.
Furthermore, MDS takes into account two categories of
documents: homogeneous sets with the same primary
context documents and heterogeneous sets with
unrelated primarycontext documents. MDS generates more
comprehensive and accurate summaries than SDS,
attempting to reconcile different and redundant
information [6].
III. Generic vs. Query-based vs. Domain-
specificGeneric-based summaries are independent of the
document and may be used by a range of end-users, while
query-based summaries are more specific summaries.

Domain-specific summaries, on the other hand, leverage
knowledge of certainfields, such as scientific and medical
publications to developmore comprehensible summaries
[7].
IV. Indicative vs. Informative
Indicative summaries contain the metadata of the text. It
gives insights of what the document is about and its main
idea. While informative summaries provide us with the
main background or domain information of the text. It
provides information about topic in an elaborated form
[4][8].
Fig -2: Types of Text Summarization
3. ABSTRACTIVE TEXT SUMMARIZATION TECHNIQUES
Methods Description Advantages Limitation
Word Graph Methodology The technique based on
word graphs is separated
into two components.
The first component is
sentence reduction,
followed by sentence
combination. This
technique involves nodes
that represent the
information about words
and their relation [9].
Word graph
technique provides
syntactically
accurate phrases
[10].
The word graph
technique creates
ungrammatical phrases
and is unconcerned with
word meaning [10].
Semantic Graph
ReductionAlgorithm
The semantic graph-based
approach builds a graph
that summarizes the
original content by
gathering semantic
information from words
and assigning weights to
nodes and edges [11].
This method's strength is
producing short,
coherent, and
grammatically accurate
phrases with few
networks [11].
This approach is
restricted to
summarizing material
from a single document
[11].
Markov
Clustering
Algorithm
To construct summaries,
the Markov Clustering
Principle employs a
hybrid technique. In this
method, sentence ranking
is accomplished by
combining linguistic
norms with the best-
fitting sentences inside a
cluster to construct new
sentences [12].
Sentences are grouped
using semantic and
statistical variables in the
Markov Clustering
Principle to produce
highly linked sentences
[12].
The accuracy of the
summary provided by
the Markov Clustering
Principle depends upon
the quality of the
sentence compression
technique [12].

Encoder-Decoder Model The encoder converts
the input sentence
sequence into a context
vector, andthe decoder
converts the processed
input into
comprehensible output
[13].
The encoder-decoder
strength is that it
addresses the vanishing
gradient issue [14].
The approach requires
an extensive dataset that
takesa long time to train
[11].
Pegasus In this approach,
significant lines are
eliminated from the input
text and compiled as
separate outputs [15].
The strength of this
method is that it selects
phrases based on
relevance rather than
randomness [16].
Pegasus may need post-
processing to remove
errors and enhance
summary text output
[17].
Summarization with
Pointer Generator
Networks
This method employs a
hybrid approach,
producing words from a
predefined vocabulary
and replicating words by
pointing [18].
This method focuses on
resolving the issue of
out- of-vocabulary
terms.
The essence of this
technique is to present
a summary based on
the source content,
rather than adding
new terminology [19].
Genetic Semantic
Graphbased Approach
The approach generates a
semantic graph from the
source text, with graph
nodes representing
predicate argument
structures (PASs) and
graph edges representing
semantic similarity
weights [20].
The merit of this method
is that it reduces
redundant information
by combining
comparable information
across documents [9].
The shortcoming of this
technique is that it fails to
recognize redundant
phrases that are
semantically similar,
resulting in an inadequate
final summary [20].
4. EXTRACTIVE TEXT SUMMARIZATION TECHNIQUES
TF-IDF Approach TF-IDF algorithm
calculates the frequency
of words in documents
and generates metric
values.
Finally, phrases with
a higher metric value
are
included in the
result[13].
The TF-IDF algorithm is
quick to compute and
has an excellent ability
to determine the
relevance of phrases
[21].
The main disadvantage
of TF-IDF is that
lengthier sentences get a
higher metric score due
to the terms' higher
occurrence in the
sentences [13].
Fuzzy Logic Fuzzy logic assigns
weights to sentences in a
document and chooses
sentences based on their
relevance, determined by
sentence length, sentence
placement, sentence
similarity, and proper
noun [27].
The advantage of fuzzy
logic is to solve the
unequal weighting of
attributes to evaluate
their relevance [30].
Fuzzy logic cannot
solve the issue of
dangling anaphora
[25].

Approach based
onClustering
The clustering technique
focuses on grouping texts
and creating cluster-level
summaries. The clusters
are generated using word
weight, sentence location,
phrase length, sentence
centrality, and proper
nouns[22].
The significance of
clustering resides in its
ability to exclude
redundant phrases from
summary automatically
[23].
The drawback of the
clustering approach is
that the summarized
phrases are not
synchronized, and
comparing the similarity
between clusters is a
challenging operation
[24].
Neural Network Approach This method works by
first training the neural
network, and then the
trained network selects
the essential phrases that
should be included in the
summary in the same
manner that a person
would [5].
The fundamental
advantage of neural
networks is their ability
to change characteristics
based on theneeds of the
user [25].
It takes an excessive
amount of time to train
aneural network [26].
Approach based
on Machine
Learning
The machine learning
technique is classified into
two types: supervised, in
which documents and
summaries are supplied,
and unsupervised, in
which just documents are
provided, and the machine
learns by evaluating them
[27].
The benefit of the
Machine Learning
technique is that it is
simple to construct and
train the model [28].
The limitation is that
significant terms often
occur in the test dataset
but not in the training
dataset are ignored [29].
5. CONCLUSION
Text summarization is a branch of Natural Language
Processing (NLP) that focuses on shortening texts and
making them more readable for users. With an excess of
data accessible on the internet and the necessity to
comprehend it in order to save the reader's time, text
summary techniques are utilized. This paper provides a
quick overview of text preprocessing, used to clean data
to do effective summarization. Then it summarizes the
many types of textsummarizing approaches, categorizing
them according to input, output, content, and purpose.
The paper's primary emphasis is on extractive and
abstractive text summarizingalgorithms based on output.
Extractive summarization summarizes by simply
extracting information from the input text. Abstractive
summarization is a more complicated method because it
summarizes the text in its language. The abstractive
technique produces better and more semantically
connected summaries. Readers would benefit
significantly from an overview of the benefits and
drawbacks of different techniques, as well as a concise
explanation. Text summarization techniques can be
applied helpfully depending on the user's needs.
6. REFERENCES
[1] Chen, J., & You, F. (2020, January). Text
Summarization Generation Based on Semantic
Similarity. In 2020 International Conference on
Intelligent Transportation, Big Data & Smart City
(ICITBS) (pp. 946-949). IEEE.
[2] Dave, H., & Jaswal, S. (2015, September). Multiple
text document summarization system using hybrid
summarization technique. In 2015 1st International
Conference on Next Generation Computing
Technologies (NGCT) (pp. 804-808). IEEE.
[3] Widyassari, A. P., Rustad, S., Shidik, G. F., Noersasongko,
E., Syukur, A., & Affandy, A. (2020). Review of
automatic text summarization techniques &
methods.Journal of King Saud University-Computer
and Information Sciences.
[4] Rajasekaran, A., & Varalakshmi, R. (2018). Review on
automatic text summarization. Inter. J. Eng. Technol, 7,
456-460.
[5] Gupta, V., & Lehal, G. S. (2010). A survey of text
summarization extractive techniques. Journal of
emerging technologies in web intelligence, 2(3),
258- 268.
[6] Kallimani, J. S. (2018, September). Survey on
extractivetext summarization methods with multi-
document datasets. In 2018 International
Conference on Advances in Computing,
Communications and Informatics (ICACCI) (pp.
2113-2119). IEEE.

[7] Boorugu, R., & Ramesh, G. (2020, July). A survey on
NLP based text summarization for summarizing
product reviews. In 2020 Second International
Conference on Inventive Research in Computing
Applications (ICIRCA) (pp. 352-356). IEEE.
[8] Gambhir, M., & Gupta, V. (2017). Recent automatic
text summarization techniques: a survey. Artificial
Intelligence Review, 47(1), 1-66.
[9] Modi, S., & Oza, R. (2018, September). Review on
Abstractive Text Summarization Techniques(ATST)
for single and multi-documents. In 2018
International Conference on Computing, Power and
Communication Technologies (GUCON) (pp. 1173-
1176). IEEE.
[10] Talukder, M. A. I., Abujar, S., Masum, A. K. M., Akter, S.,
&Hossain, S. A. (2020, July). Comparative Study on
Abstractive Text Summarization. In 2020 11th
International Conference on Computing,
Communication and Networking Technologies
(ICCCNT) (pp. 1-4). IEEE
[11] Mridha, M. F., Lima, A. A., Nur, K., Das, S. C., Hasan, M.,
&Kabir, M. M. (2021). A Survey of Automatic Text
Summarization: Progress, Processand Challenges.
IEEEAccess, 9, 156043-156070
[12] Sahoo, D., Bhoi, A., & Balabantaray, R. C. (2018).
Hybrid approach to abstractive summarization.
Procediacomputer science, 132, 1228-1237.
[13] Shinde, M., Mhatre, D., & Marwal, G. (2021, March).
Techniques and Research in Text Summarization-A
Survey. In 2021 International Conference on
Advance Computing and Innovative Technologies in
Engineering (ICACITE) (pp. 260-263). IEEE
[14] Wikipedia Contributors. (2020). Multi-
DocumentSummarization— Wikipedia, the Free
Encyclopedia. Accessed: Oct. 8, 2021.
[Online]. Available:
https://en.wikipedia.org/w/index.php?title=Multi-
document_ summarization%&oldid=986613170
[15] Gupta, A., Chugh, D., & Katarya, R. (2021). Automated
News Summarization Using Transformers. arXiv
preprint arXiv:2108.01064
[16] Zhang, J., Zhao, Y., Saleh, M., & Liu, P. (2020, November).
Pegasus: Pre-training with extracted gap-sentences for
abstractive summarization. In International
Conference on Machine Learning (pp. 11328-
11339). PMLR.
[17] Oliveira, L. M. R., Busson, A. J. G., Carlos de Salles, S.
N., dos Santos, G. N., & Colcher, S. (2021,
November). Automatic Generation of Learning
Objects Using Text Summarizer Based on Deep
Learning Models. In Anais do XXXII Simpósio
Brasileiro de Informática na Educação (pp. 728-
736). SBC.
[18] Anh, D. T., & Trang, N. T. T. (2019, December).
Abstractive text summarization using pointer-
generator networks with pre-trained wordembedding.
In Proceedings of the tenth international
symposium on information and communication
technology (pp. 473-478).
[19] Boutkan, F., Ranzijn, J., Rau, D., & van der Wel, E. (2019).
Point-less: More abstractive summarization with
pointer-generator networks. arXiv preprint
arXiv:1905.01975.
[20] Khan, A., Salim, N., & Kumar, Y. J. (2015, October).
Genetic semantic graph approach for multi-
document abstractive summarization. In 2015 Fifth
International Conference on Digital Information
Processing and Communications (ICDIPC) (pp. 173-
181). IEEE.
[21]Allahyari, M., Pouriyeh, S., Assefi, M., Safaei, S., Trippe, E.
D., Gutierrez, J. B., & Kochut, K. (2017). Text
summarization techniques: a brief survey. arXiv
preprint arXiv:1707.02268.
[22]Deshpande, A. R., & Lobo, L. M. R. J. (2013). Text
summarization using clustering technique.
International Journal of Engineering Trends and
Technology, 4(8), 3348-3351.
[23]Jewani, K., Damankar, O., Janyani, N., Mhatre, D., &
Gangwani, S. (2021, March). A Brief Study on
Approaches for Extractive Summarization. In 2021
International Conference on Artificial Intelligence
andSmart Systems (ICAIS) (pp. 601-608). IEEE.
[24]Akter, S., Asa, A. S., Uddin, M. P., Hossain, M. D., Roy, S.
K., & Afjal, M. I. (2017, February). An extractive text
summarization technique for Bengali document (s)
using K-means clustering algorithm. In 2017 IEEE
International Conference on Imaging, Vision &
PatternRecognition (icIVPR) (pp. 1-6). IEEE.
[25]Moratanch, N., & Chitrakala, S. (2017, January). A
survey on extractive text summarization. In 2017
international conference on computer,
communication and signal processing (ICCCSP) (pp.
1-6). IEEE.
[26]Andhale, N., & Bewoor, L. A. (2016, August). An
overviewof text summarization techniques. In 2016
International Conference on Computing
Communication Control and automation (ICCUBEA)
(pp. 1-7). IEEE.
[27]Kumar, A. K. S. H. I., & Sharma, A. D. I. T. I. (2019).
Systematic literature review of fuzzy logic based
text summarization. Iranian journal of fuzzy systems,
16(5),45-59.
[28]Patel, R., Thakkar, A., Makwana, K., & Patel, J. (2017,
March). Comprehensive and Evolution Study
Focusing on Comparative Analysis of Automatic
TextSummarization. In International Conference on
Information and Communication Technology for
Intelligent Systems (pp. 383-389). Springer, Cham.
[29]Lagrini, S., Redjimi, M., & Azizi, N. (2017). Automatic
arabic text summarization approaches.
International Journal of Computer Applications,
164(5), 31-37.
[30]Babar, M. S. (2014). Improving Text Summarization
Using Fuzzy Logic (Doctoral dissertation,
RAJARAMBAPU INSTITUTE OF TECHNOLOGY.

A Survey on Automatic Text Summarization

Recommended

More Related Content

Similar to A Survey on Automatic Text Summarization (20)

More from IRJET Journal (20)

Recently uploaded (20)

A Survey on Automatic Text Summarization