Offline vs. Online Sentiment Analysis: Issues With Sentiment Analysis of Online Micro-Texts

Offline vs. Online Sentiment Analysis: Issues With Sentiment Analysis of Online Micro-Texts

Ritesh Srivastava (University of Delhi, New Delhi, India) and M.P.S. Bhatia (University of Delhi, New Delhi, India)
Copyright: © 2017 |Pages: 18
DOI: 10.4018/IJIRR.2017100101
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

Recently, the social networking sites (SNSs) have proven their immense power of prediction for predicting the results of the real-world events. However, for real-time monitoring of the world activities via microblogging site like Twitter, it is important to perform the sentiment analysis of online micro-texts in real-time to support fast and intelligent decision-making and hence to execute the appropriate actions in the real world in real-time. In this context, this paper discusses the online sentiment analysis process of online micro-texts in perspectives of the real-time analysis process. In addition, this paper argues the non-applicability of the classical time consuming Natural Language Processing (NLP) methods and the affinity of Machine Learning (ML) methods in performing the online sentiment analysis by contrasting it with offline sentiment analysis. Furthermore, it also formalized the online sentiment analysis process of online micro-texts by raising novel issues and proposing new performance measures for online sentiment analysis.
Article Preview

The problem of automatic determination of the semantic orientation of words and documents by their polarities is not new in the community of linguistic researchers. A pioneered approach for clustering of adjectives by their polarities has been described in (Hatzivassiloglou & McKeown, 1997), which uses contrary and cooperating conjuncts for this task. Various machine learning methods like Support Vector Machine (SVM) (Joachims, 1998), and Naïve Bayesian (NB) have gained interest for performing automatic text classification in the late 1990s (Agarwal & Mittal, 2014) for organizing the e-documents in online digital libraries. Till that time, the targeted e-documents include stories, movies, books, technical papers, etc. These e-documents are usually very consistent on grammar and spelling. However, with the boom in e-commerce sites an extreme affection of linguistic researchers has been observed in classifying the online reviews of products on the basis of writer’s opinion on a product into positive and negative polarities (Ding et al., 2008; Turney, 2002; Liu & Cheng, 2005; Srivastava et al., 2010; Baumgarten et al., 2013), and a new terminology i.e. “Opinion–Mining” (OM) has been coined in the horizon of Natural language processing (NLP) and text mining for collective representation of the works of similar kind (Ohmura, et al., 2014; Hariharan & Ramkumar, 2011). Up to this era of time, most of the tasks of sentiment analysis favor some hybrid approaches using both: the time-consuming classical methods of NLP (e.g. Rule-based) and Statistical and Machine Learning (ML) approaches.

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 8: 4 Issues (2018): 1 Released, 3 Forthcoming
Volume 7: 4 Issues (2017)
Volume 6: 4 Issues (2016)
Volume 5: 4 Issues (2015)
Volume 4: 4 Issues (2014)
Volume 3: 4 Issues (2013)
Volume 2: 4 Issues (2012)
Volume 1: 4 Issues (2011)
View Complete Journal Contents Listing