Sentiment Analysis using Optimized Feature Sets in Different Twitter Dataset Domains
Anup Raut1, Rahul K. Pandey2
1Anup Raut, PhD Research Student, MUIT, Lucknow.
2Dr. Rahul K. Pandey, Research Guide, MUIT, Lucknow.
Manuscript received on 24 August 2019. | Revised Manuscript received on 02 September 2019. | Manuscript published on 30 September 2019. | PP: 3035-3039 | Volume-8 Issue-11, September 2019. | Retrieval Number: K21950981119/2019©BEIESP | DOI: 10.35940/ijitee.K2195.0981119
Open Access | Ethics and 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: In the last years, the relevance of sentiment analysis is broad and dominant. The capability to take out insights from social data is a tradition that is being extensively accepted by all over globe. Sentiment Analysis has turn out to be a hot-trend issue of technical and marketplace research in the area of Natural Language Processing (NLP) and Machine Learning. Sentiment analysis is enormously useful in social media supervising as it permits us to expand an impression of the wider open estimation behind definite topics. Investigation of social media streams is typically limited to just essential sentiment analysis and count based metrics. This is of the same kind to just scratching the outside and missing out on those elevated value insight that is ahead of you to be discovered. There’s a lot of effort to be done, but perfections are being prepared every day. It is a way to appraise on paper or verbal language to settle on if the expression is favorable, unfavorable, or unbiased, and to what level. Today’s algorithm-based sentiment analysis tools can touch vast amount of client response constantly and precisely. Balancing with text analytics, sentiment analysis exposes the customer’s estimation concerning topics ranging from your goods and services to your position, your advertisements, or even your challengers. These efforts scrutinize the crisis of studying texts, like posts and reviews, uploaded by user on Twitter. The Support Vector Machine (SVM), k-nearest neighbors algorithm (KNN) and proposed optimized feature sets model is offered to progression the tweet features and to recognize the out of sight sentiments from these tweets. These essential concepts when used in combinations become a very significant tool for analyzing millions of variety conversations with human echelon accurateness. The projected optimized feature sets model Sentiment Analysis exercise the assessment metrics of Precision, Recall, F-score, and Accuracy. Also, average measures weighted F1-scores are constructive for categorization of Positive, Negative and Neutral multi-class problems. The running time of the technique is evaluates by accomplishing diverse methods in the same investigational setup consisting a cluster of 8 nodes. Planned optimized feature sets model Sentiment Analysis reachs 82 % accuracy as compare with SVM 78.6 % and KNN 75 %. Further, while analyzing sentiments of tweets we have measured only tweets in English acknowledged by Twitter streaming API.
Keywords: NLP, SVM, KNN,F1-score.
Scope of the Article: Software Domain Modelling and Analysis.