期刊名称：International Journal of Innovative Research in Computer and Communication Engineering
印刷版ISSN：2320-9798
电子版ISSN：2320-9801
出版年度：2018
卷号：6
期号：6
页码：6230-6235
DOI：10.15680/IJIRCCE.2018.0606019
出版社：S&S Publications
摘要：Twitter is a popular microblogging service where users create tweets which sometimes express opinions
about different topics. Sentiment analysis of twitter data is useful for companies that want to monitor the public
sentiment of their brands also for consumers who want to research the sentiment of products before purchase. Existing
approaches has high complexities, less throughput and consumes more computation time for large dataset. In this
paper, we will find polarity of tweets using lazy learning method- K-nearest neighbors classifier. For classification and
representing text data when modeling with machine learning we use bag-of-words model here, as it is simple to
understand and implement. We performed our experiments using 1.6 million tweets. In order to manage the
preprocessed data we are using corpus. Our Experimental evaluations show that our proposed technique is efficient,
provides maximum throughput and consumes less time as compared to previous works. We achieve accuracy of 81.1%
and specificity of 84.7% of test dataset which is much better than existing approaches.
