A web based approach: Acronym Definition Extraction

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 03 Issue: 02 | Feb-2016 www.irjet.net p-ISSN: 2395-0072
© 2016, IRJET | Impact Factor value: 4.45 | ISO 9001:2008 Certified Journal | Page 1199
A web based approach: Acronym Definition Extraction
R. Menaha*, M.Barkavi , P.Guha Prashanthini , R.Narmadha
*Assistant Professor, Information Technology, Dr. MCET, Tamil Nadu, India
B.Tech - Information Technology, Dr. MCET, Tamil Nadu, India
Abstract
Acronyms are widely used in the tasks like web searches, tweets, text messages, and mails etc. Acronyms are typically
ambiguous and often disambiguated by context words. Due to its dynamicity, different approaches has been experimented in
the past decade to extract acronym definition. Different manual edited websites like acronym finder, acronyms. the free
dictionary.com, nalingo.com is accessible to list the definition of an acronym. However an automated method is required to
support this identification. This paper presents an automatic web based approach to extract the definitions of an acronym.
The proposed system uses the web resources of Google, Wikipedia, Bing and Acronym Finder to identify the definitions of an
acronym. From Google and Bing web pages, the snippets and titles are extracted and pattern extraction algorithm is applied to
identify the definitions. Similarly the Wikipedia and acronym finder web pages are extracted to find the acronym definitions.
The extracted definitions might be applied in information retrieval, question answering system and query expansion area as
future work.
Keywords: Acronym definition extraction, Abbreviation extraction
-----------------------------------------------------------***----------------------------------------------------------------
1. INTRODUCTION
Acronyms are abbreviations formed from the initial components of words or phrases. Acronyms are textual forms used to
stress the importance of entities and provide an alternative way to refer the same entity which is easier to understand .Some
of the characteristics of acronyms are:
1. Dynamicity: New Acronyms are defined in every domain and every day. This is evident in social networks like
twitter, LinkedIn, online chat.
2. Ambiguity: Each acronym has different meanings in different domain [e.g., SOAP has 35 definitions in acronym finder
and 32 definitions in acronyms.thefreedictionary.com]. This can be disambiguated by using context words [e.g., SOAP XML
refers Simple Object Access Protocol, SOAP Analysis refers Spectrometric Oil Analysis Program].
3. Diverse of Generality: Some acronym –definition pairs are commonly used, and some of them are very rare [e.g.,
NASA –National Aeronautics and space Administration [most common], NASA [Newspaper Advertising Sales Corporation].
Usage of acronyms is more in the applications like Biomedicine, Natural language Processing, Ontology population,
question answering system and web search. Due to its dynamicity, it is difficult to maintain up-to-date lexical
repository of all the acronyms and their meanings. Some of the manually edited websites like acronym finder,
acronyms.thefreedictionary.com have millions of acronyms and meanings. Automatic discovery of acronym
definitions also attempted in the past decade most of them are language dependent. This paper proposes an
automated web based approach to identify the definitions of an acronym. It uses the web resources of google, Bing,
Wikipedia, and acronym finder. The system extracts the snippets, titles and apply pattern extraction algorithm to
discover the list of definitions. The rest of the paper is organized as follows. Section 2 defines previous works in the area of
acronym definition discovery. Section 3 describes in detail the proposed methodology. Section 4 presents the results and
evaluation, showing the obtained results and discussing them. The last section gives some conclusions and proposes some
lines of future work.

2. RELATED WORKS
Andr’e Kempe [1], the report shown how the alignment-based approach to acronym-meaning extraction can be implemented
by means of a 3-tape weighted finite-state machine (3-WFSM). The 3-WFSM will read a text chunk on tape 1 and an acronym
on tape 2, and generate all possible alignments on tape 3, inserting dots to mark which letters are used in the acronym.
CvetanaKrstev, DuskoVitas, RankaStankovi’c[3] presented a comprehensive approach to acronyms for Natural-Language
Processing (NLP) of Serbian texts. The procedure includes extraction of acronyms and their definitions that are usual Multi-
Word Units (MWUs), shallow parsing of MWUs that enables MWU lemmatization and production of entries in morphological
electronic dictionaries, both for MWU and acronyms, that are provided with grammatical, syntactic, semantic and domain
information
Dana Dannels [4], the approach is based on that acronym-definition pairs follow a set of patterns and other regularities that
can be usefully applied for the acronym identification task. Supervised machine learning was applied to monitor the
performance of the rule-based method, using Memory Based Learning (MBL).
David Sanchez, David Isren [5], proposed methodology which has been divided into several part, first a simple algorithm
generates the possible acronym by the combination of alpha numeric characters of a specific length. For all those generated
candidates, the system tries to discover all possible definitions from the web. Finally the list of definitions found is filtered
using a set of general rules.
Jun Xu.Yalou Huang [7], proposed a novel machine learning approach to extract acronyms from text. First, all likely acronyms
are identified by heuristic rules. Second, expansion candidates against each likely acronym are generated from the
surrounding text. He used the support vector machine (SVM) model to select the genuine expansions for acronyms.
Sunghwan sohn, Donald C Comeau, Won Kim and W John Wilbur [10], proposed an abbreviation identification algorithm that
employs a number of rules to extract potential SF-LF pairs and variety of strategies to identify the most probable LFs.The
reliability of the strategy can be estimated which they term pseudo – precision (P-precision).
3. METHODS
The list of acronyms are extracted from four different web resources. They are Google, Bing, Wikipedia and Acronym Finder.
The method that are involved in the expansion identification differs for different web resources. The outline of proposed
system implementation is given in fig [1.0]
Fig 1.0 Outline of system implementatio

3.1 Google and Bing Web pages
Google and Bing pages contains snippets, titles, URL and other information sources. Usually the snippets and titles contains the
acronym definitions of given input acronym query. Hence the snippets and titles are extracted from Google and Bing web
pages they are stored in a local repository. Pattern extraction algorithm presented in fig 5.0 is applied to extract the patterns
from snippets and titles.
3.1.1 Snippet Extraction
The snippet contains the small hint of information that the user request from the search engine. Snippets are useful for
searches because most of the time, the user can read the snippet and decide whether particular search result is relevant
without opening the URL. The processing of snippet is also efficient because it prevent from errors in downloading web pages,
which might be time consuming depending on the size of the web pages. This extracted snippet is stored in the local database,
once the new data comes existing data gets overridden.
Fig 2.0 Snippet extracted from google for CPU
3.1.2 Title Extraction
Web pages contains the title which shows the detail that are available in the document may contains the definitions of an
acronym.
Fig 3.0 Title extracted from google
A central processing unit (CPU) is the electronic
circuitry within a computer that carries out the
instructions of a computer program by
performing the basic…
Central processing unit - Wikipedia, the free
encyclopedia
Images for CPU
What is Central Processing Unit (CPU)?
Webopedia

Fig 4.0 Outline of System Implementation
3.2 Wikipedia
Wikipedia follows the procrastination principle regarding the security of its content. It started almost entirely open—anyone
could create articles, and any Wikipedia article could be edited by any reader, even those who did not have a Wikipedia
account. Modifications to all articles would be published immediately. Whenever an acronym is searched through Wikipedia, it
retrieve the results in various contexts and meaning using fig.5.0. The pages can be stored in a local repository to extract
snippets and titles. In Wikipedia the web pages gets extracted, and the Acronym definitions are extracted from the
disambiguous webpages. The patterns for input acronym query is retrieved using the pattern extraction pseudocode given in
fig.5.0
3.3 Acronym Finder
The web pages of Acronym Finder contains the expansion for input acronym. The Acronym Finder would contain five domains
like Science and Medicine, Business, Military and Government, Organizations and Information Technology. First of all, the user
would be asked to select the domain and then the acronym must be given. To extract web pages, an acronym is given as input
to the search engine and the results of search engine are stored in local database. From the extracted web pages, the
expansions are fetched using pattern extraction procedure using fig.5.0 and stored in lexical repository. In acronym finder the
web pages are extracted, which contains the list of acronym expansion from different domains. Based on the user interest the
corresponding meaning of an acronym in its domain are listed.

3.4 Patterns Extraction
A Pattern is a series of action or event or behavior that together show how things normally happen or done. The following
pattern extraction algorithm fig. 5.0 is devised to extract the acronym definitions from snippets and titles of Google, Bing,
Wikipedia and Acronym finder.
Input: Acronym
Output: List of Acronym defifntions
/* Google, Bing, Wikipedia and Acronymfinder web pages are used for the acronym definition extraction */
/* Google and Bing */
do
For each acronym
a. α1  Extract the google web pages
b. S1 Extract the snippets from α
c. T Extract the titles from α
d. P1 Extract the patterns from S and T. end
/* Wikipedia */
do
For each acronym
a. α2  Extract the Wikipedia disambiguation web pages
b. S2 Extract the required snippets from α
c. P2  Extract the patterns from S2.
end
/* Acronym Finder */
do
For each acronym
a. α3  Extract the Wikipedia disambiguation web pages
b. S3 Extract the required snippets from α
c. P3  Extract the patterns from S3.
End.
/*** To extract patterns Pattern Extraction algorithm is applied ***/

Fig 5.0 Pattern Extraction Pesudocode.
Input: Snippets /Titles
Output : Lists of Acronyms
Variables used: L,S ,F, T flag;
1. L Find the length of the Acronym
2. S [0-L] Store each character of given acronym in an array.
3. Read the snippets / titles.
4. Replace all the numbers [0-9] and Special symbols [, “ . ? etc.,] with empty space.
5. Remove stop words from the snippets and titles. [e.g., a,an, the , between , etc.,]
6. F Read the snippet / title file and store it into buffer
While (F! =eof)
{ a. Read line from buffer F and apply tokenization.
b. T Read token
c. If(T!==NULL)
{ for (i=0;i<l;i++)
T[i] Read the first character of token
if (T[i]==S[i])
{ flag++;
add the token into wordlist
goto step b.
}
else
goto step b.
if (flag==l)
add that wordlist into pattern file.
}
}

5. RESULTS AND DISCUSSION
Thus the expansions of acronyms have been extracted from four web resources like Acronym Finder, Bing and Google
Pages and Wikipedia. For experimental purpose 100 acronym definitions are extracted from that 4 web resources. As an
example 5 acronyms and extracted definitions from acronym finder, Bing, Google and Wikipedia is listed in table 1.0. For
implementation, Java, JSON, Swing and Google search engine API were used. The extracted definitions will be applied in
query expansion system as future work.
6. CONCLUSION
Acronyms are widely used in various applications. This paper proposed a method on extracting acronym definitions from
Google, Bing, Wikipedia and acronym finder. Acronym definition extraction for about hundred acronyms is extracted
successfully from the various web resources. As a future work extracted definitions would be applied in query expansion
for effective information retrieval.
Acronym Acronym Finder Bing and Google Wikipedia
CPU
Central Processing Unit
Communist Party of Ukraine
Chemical Production Unit
Central Philippine University
Commonwealth Press Union
Central Policy Unit
Computer Power Use
Cost Per Unit
Call Pick Up
Critical Path Update
Computer Processing Unit
Common Party User
Cost Per Unit
Columbia Pacific University
Central Philippine University
Commonwealth Press Union
Clark Public Utilities
Columbia Pacific University
China Pharmaceutical University
Chemical Production Unit
SOAP
Subjective Objective Assessment Plan
Simple Object Access Protocol
Seal Of Approval Process
Symbolic Optimal Assembly Program
Spectrometric Oil Analysis Program
Summary On A Page
Small Operator Assistance Program
Students Organized Against Prejudice
Students Organized Against Poverty
Summary On A Page
Society Of Airway Pioneers
Strategy On A Page
Seal Of Approval Process
Symbolic Optimal Assembly
Program
Spectrometric Oil Analysis
Program
Snakes On A Plane
Students Organized Against
Prejudice
Students Organized Against
Poverty
NSS
National Security Strategy
Network Switching Subsystem
National Service Scheme
Names Service Switch
National Service Switch
Network Security Scanner
Nadal Switching System
Naval Surface Strike
National Security Service
Novel Storage Services
Not So Sure
National Shelter System
National Security Service
National Service Scheme
National Scheme Services
New Skies Satellites
Nair Service society
NBA
National Basketball Association
National Blood Association
Next Best Alternative
National Band Association
Narmada Bachao Andolan
National Book Award
National Boxing Association
National Business Association
National Book Award
No Boys Allowed
Network Behavior Analysis
Netbook Book Agreement
N-Butyl Alcohol
Narmada Bachao Andolan
National Book Award
National Boxing Association
Nepal Basketball Association
National Braille Association
Table 1.0 Extracted definitions for CPU, SOAP, NSS and NBA from Google, Bing, Wikipedia and Acronym Finder.

REFERENCE
[1] Andr´e Kempe, “Acronym-Meaning Extraction from Corpora Using Multitape Weighted Finite-State Machines”, Springer Verlag,
arxiv.org, Dec 2006, cs/0612033
[2].Bilyana Taneva1∗, Tao Cheng2, Kaushik Chakrabarti2, Yeye He “Mining Acronym Expansions and Their Meanings Using Query Click
Log”, WWW 2013, May 13–17, 2013, Rio de Janeiro, Brazil. ACM 978-1-4503-2035-1/13/05.
[3] Cvetana Krstev, Duško Vitas, Ranka Stankovi´ , “A Lexical Approach to Acronyms and their Definitions”
[4]. Dana Dannels, “Automatic Acronym Recognition”, ACM DL, EACL ’06, Pg:167-170
[5] David Sanchez, David Isren, “Automatic extraction of acronym definitions from theWeb” Apple Intell(2011) 34: 311-327 DOI
10.1007/s 10489-009-0197-4
[6] Jain A, Cuzerzan S, Azzam S, “Acronym –Expansion Recognition and Ranking on the web, IEEE International conference on
Information Reuse and Integration 2007, pg:209-214. DOI: 10.1109/IRI.2007.4296622.
[7] Jun Xu.Yalou Huang, “using SVM to extract acronyms from text”, Springer Verlag 2006, DOI 10.1007/s00500-006-0091-5.
[8] Leah S. Larkey, Paul Ogilvie, M. Andrew Price, Brenden Tamilio, “Acrophile: An Automated Acronym Extractor and Server”, ACM 2000,
Pg:205-214, DOI:10.1145/336597.336664
[9] Min song, Peishih Chang, “Automatic Extraction of Abbreviation for Emergency Management Websites”,ISCRAM conference- Washington
May 2008, Pg:93-100.
[10] Sunghwan sohn, Donald C Comeau, Won Kim and W John Wilbur, Abbreviation definition identification based on automatic
precision estimates, BMC Bioinformatics 2008, 9:402 doi:10.1186/1471-2105-9-402. Sep 2008.
[11] www.acronymfinder.com

A web based approach: Acronym Definition Extraction

More Related Content

What's hot (20)

Similar to A web based approach: Acronym Definition Extraction (20)

More from IRJET Journal (20)

Recently uploaded (20)

A web based approach: Acronym Definition Extraction