Domain Specific Lexicon Generation through Sentiment Analysis

Authors

  • Kamran Shaukat The University of Newcastle, Australia
  • Ibrahim A Hameed Norwegian University of Science and Technology, Norway
  • Suhuai Luo The University of Newcastle, Australia
  • Imran Javed University of the Punjab, Pakistan
  • Farhat Iqbal University of the Punjab, Pakistan
  • Amber Faisal University of the Punjab, Pakistan
  • Rabia Masood University of the Punjab, Pakistan
  • Ayesha Usman University of the Punjab, Pakistan
  • Usman Shaukat University of the Punjab, Pakistan
  • Rosheen Hassan University of the Punjab, Pakistan
  • Aliya Younas University of the Punjab, Pakistan
  • Shamshair Ali University of the Punjab, Pakistan
  • Ghazif Adeem University of the Punjab, Pakistan

DOI:

https://doi.org/10.3991/ijet.v15i09.13109

Keywords:

domain specific, General-language dictionary, lexicon, polarity, Sentiment analysis

Abstract


Sentiment analysis (SA) is used to extract opinions from a huge amount of data and these opinions are comprised of multiple words. Some words have different semantic meanings in different fields and we call them domain specific (DS) words. A domain is defined as a special area in which a collection of queries about a specific topic are held when user do queries in the data regarding the domain appear. But Single word can be interpreted in many ways based on its context-dependency. Demonstrate each word under its domain is extremely important because their meanings differ from each other so much in different domains that a word meaning from A in one context can change into Z in another context or domain. The purpose of this research is to discover the correct sentiment in the message or comment and evaluate it either it is positive, negative or neutral. We collected tweets dataset from different domains and analyze it to extract words that have a different definition in those specific domains as if they are used in other fields of life they would be defined differently. We analyzed 52115 words for finding their DS meaning in seven different domains. Polarity had been given to words of the dataset according to their domains and based on this polarity they have been recognized as positive negative and neutral and evaluated as domain-specific words. The automatic way is used to extract the words of the domain as we integrated and afterward the comparison to identify that either this word differs from other words as far as domain is concerned. This research contribution is a prototype that processes your data and extracts their domain-specific words automatically. This research improved the knowledge about the context-dependency and found the core-specific meanings of words in multiple fields.

Downloads

Published

2020-05-15

How to Cite

Shaukat, K., A Hameed, I., Luo, S., Javed, I., Iqbal, F., Faisal, A., Masood, R., Usman, A., Shaukat, U., Hassan, R., Younas, A., Ali, S., & Adeem, G. (2020). Domain Specific Lexicon Generation through Sentiment Analysis. International Journal of Emerging Technologies in Learning (iJET), 15(09), pp. 190–204. https://doi.org/10.3991/ijet.v15i09.13109

Issue

Section

Papers