DFWC: Discriminative Feature Weight Correlation based Sentiment Analysis of Social Networks
Muddada Muralikrishna1, G. Lavanya Devi2
1Muddada Muralikrishna, Research Scholar, Department of Computer Science and Systems Engineering, AUCE (A), Andhra University, Visakhapatnam (Andhra Pradesh), India.
2Dr. G. Lavanya Devi, Department of Computer Science and Systems Engineering, AUCE (A), Andhra University, Visakhapatnam (Andhra Pradesh), India.
Manuscript received on 27 March 2019 | Revised Manuscript received on 06 April 2019 | Manuscript Published on 27 April 2019 | PP: 706-716 | Volume-7 Issue-6S2 April 2019 | Retrieval Number: F11210476S219/2019©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Sentiment analysis is the buzz in current era of the social network-based opinion sharing about products, events, politics and many more. The lexicon analysis and machine learning are the major strategies to perform sentiment analysis. The phenomenal escalation of available data in social networks is indeed evincing the intensification of dimensions inportraying the opinion. Hence, the machine learning is the much effective than the lexicons analysis. However, the escalation of the dimensions in projection of the sentiment, indicating the need of the optimal learning methods enables to learn from the correlation between the features of vivid dimensions, which is the crucial requirement for available labeled data with multidimensional features. This manuscript portrayed a novel method that learns from the discriminative feature weight correlation to train the classifier in regard to identify the polarity of the sentiment to the given opinion about the target event. In this regard, the feature weights are used to identify the impact of the corresponding feature towards positive or negative polarity. The experimental study evincing the performance advantage of the proposal that compared to other contemporary models.
Keywords: Sentiment Analysis, Query Expansion Ranking (QER), NLP, Feature Selection Strategy.
Scope of the Article: Social Networks