Analisis Sentimen Performansi Operator Telekomunikasi di Indonesia Menggunakan Metode Text Mining

Abstract

The telecommunications sector in Indonesia has experienced rapid development in recent years, characterized by the increasing number of telecommunications operators offering various services and products. Therefore, there is competitive rivalry among operators. The right strategy is needed to survive and compete effectively. One of the efforts that can be made by telecommunication companies is to evaluate operational performance. This research aims to analyze the sentiment of X users towards telecommunication operational performance, at Telkomsel and Tri operators using text mining methods, namely Naïve Bayes, Support Vector Machine (SVM) and Decision Tree Learning (DTL). The research data is obtained by crawling from the X application, then the data is processed to remove unnecessary words or affixes. Then data modeling and validation is carried out using split validation and cross validation techniques. In the split validation technique, the data is divided into 70% training data and 30% testing data, while in the cross-validation technique the fold parameter is set to determine which fold has the highest accuracy. The results of the study show that the SVM method has the highest accuracy, where in split validation the accuracy is 84.29% for Telkomsel data and 75.70% for Tri data. Similarly, in cross validation, the accuracy is 82.15% on fold 4, 7 for Telkomsel data and 61.41% on fold 9 for Tri data. In addition, it is known that Telkomsel data has 18.64% positive sentiment and 81.36% negative sentiment. While Tri data has 61.11% positive sentiment and 38.89% negative sentiment.