An Ensemble Sentiment Classification System of Twitter Data for Airline Services Analysis: Smart Use of Soft Computing

To perform the airline services analysis it is difficult to gather the data by means of questionnaire investigation as it is both labour and time consuming. Using information obtained from Twitter is a useful method for gathering customer feedback and enhancing services. Everyone uses social media today, therefore there are a lot of user perspectives available. Twitter social media data frequently employs confusing phrasing that makes it challenging to distinguish between positive and negative sentiments. More than a billion social media communications need to be properly kept in a database and analysed in order to interpret the data. In this study, the majority vote principle was used to guide the use of an ensemble sentiment classification technique that combined the C4.5 Decision Tree, SVM, Bayesian Network, Naive Bayes and Random Forest algorithms. The results show that the suggested ensemble classifier performs better than both majority voting ensemble classifiers and stand-alone classifiers. According to our findings, the ensemble approach might increase overall mining and sentiment classification performance for other services as well.