Sentiment analysis of prospective buyers of mega online sale using tweets

Client sentiments assume an extremely significant part of daily life. Whenever somebody needs to take a choice, they tend to consider the feelings of others. Since there is a decent measure of social information present on the Web consequently, it is vital to utilize techniques that automatically classify them. This paper analyses the sentiments of prospective buyers of mega online sell using their posted tweets about the big billions day sale. Experiments are performed using twenty-one thousand randomly extracted tweets using #Big Billion Day hashtag. Three individual learners such as Naive Bayes, maximum entropy classification, and support vector machine (SVM) are used with unigram and bigram features for classifying text sentiments (tweets) as positive or negative. The learners' performance is evaluated using precision, recall, and F-measure performance metrics. The results suggest that learners with bigram features produce better results as compared to unigram features, and general sentiments are positive for this mega sale.