Evaluation of Customer Satisfaction about Telecom Operators in Turkey by Analyzing Sentiments of Customers through Twitter

Sentiment analysis is the process of determining whether a piece of writing is positive, negative or neutral. Text analysis based sentiment analysis consolidates natural language processing models and machine learning techniques to determine sentiment scores to the entities, topics, themes and categories within a phrase or, sentence. Furthermore, customer satisfaction is an evaluation of how products and services supplied by a company satisfy or exceed customer expectation. In this work, we propose to analyze customer satisfaction of three big telecommunication operators which are Turkcell, Turk Telekom, and Vodafone in Turkey by utilizing sentiment analysis of customers of them. For this purpose, Twitter social media platform is used for the purpose of gathering the related tweets that are mentioned with hashtags by the customers of operators. In order to improve the system performance, various pre-processing models are used such as removing punctuation marks, stop-words elimination, removing tags, URLs filter, stemming. Finally, sentiment of users is evaluated through machine learning algorithms namely, random forest, support vector machine (SVM), multilayer perceptron (MLP), k-nearest neighbors (k-NN), naive Bayes (NB), and decision tree. The experiment results present remarkable classification performance with accuracy of over 80 percent for all telecom operators. Thus, this study can inspire telecommunications companies to analyze customer satisfaction through the social media platform.