Analisis Sentimen pada PLN Mobile Menggunakan Metode Decision Tree

Abstract

PLN (Persero) adalah perusahaan kelistrikan yang menyediakan layanan kepada masyarakat Indonesia. Mereka telah mengembangkan aplikasi bernama PLN Mobile, yang memungkinkan pelanggan untuk mengakses informasi dan layanan terkait listrik dengan mudah. Dalam konteks penggunaan media sosial, analisis sentimen menggunakan metode Decision Tree digunakan untuk mengklasifikasikan tweet-tweet yang berasal dari twitter dan berkaitan dengan Aplikasi PLN Mobile menjadi kategori sentimen positif, negatif, atau netral. Proses analisis sentimen dimulai dengan mengumpulkan tweet-tweet yang relevan dan memisa hkannya ke dalam kelas sentimen yang sesuai. Fitur-fitur penting seperti kata-kata kunci dan pola diekstraksi dari tweet-tweet tersebut. Hasil penelitian menunjukkan bahwa metode Decision Tree mencapai akurasi tertinggi sebesar 96,5%, presisi sebesar 87,2%, dan recall sebesar 83%. Analisis sentimen terhadap tanggapan masyarakat terhadap PLN Mobile mengungkapkan bahwa kata-kata kunci yang muncul dalam sentimen positif adalah "mudah", "semua", dan "terima". Sentimen negatif mengungkapkan dua masalah utama, yaitu akses dan permasalahan. Sentimen netral melibatkan tiga topik, yaitu layanan, melalui, dan dapat. Untuk penelitian selanjutnya, disarankan untuk menggunakan metode atau algoritma lain sebagai perbandingan guna meningkatkan sistem analisis sentimen. Hal ini akan memberikan pemahaman yang lebih komprehensif tentang sentimen pengguna terhadap Aplikasi PLN Mobile dan memberikan wawasan berharga bagi PLN dalam pengembangan dan perbaikan layanan mereka. PLN (Persero) is an electricity company that provides services to the people of Indonesia. They manage power plants and transmit electricity throughout the country. PLN continues to innovate by creating an application called PLN Mobile, which is an Android-based application. This application is integrated with the Integrated Complaints and Complaints Application (APKT) and the Centralized Customer Service Application (AP2T). With PLN Mobile, customers can easily access information about electricity bills, adding power, submitting complaints, and other information related to electricity services. Twitter, as a popular social media in Indonesia, is used by many people to interact and share opinions in the form of tweets. Through tweets on Twitter, users can discuss and provide opinions about the PLN Mobile App. Sentiment analysis using the Decision Tree method aims to classify tweets related to the PLN Mobile Application into positive, negative, or neutral sentiment categories. The Decision Tree method uses machine learning algorithms to build a decision tree based on features present in the tweet dataset. The sentiment analysis process with the Decision Tree method begins with a data pre-processing stage, where tweets relevant to the PLN Mobile Application are collected and separated into the appropriate sentiment classes (positive, negative, or neutral). Next, relevant features such as key words, phrases, or patterns will be extracted from the tweets. Based on the results of the research that has been done, sentiment analysis on PLN mobile using the decision tree method has been successfully carried out with the highest accuracy of 96.5%, precision of 87.2%, and recall of 83%. Sentiment analysis for public responses to PLN mobile has words that appear in positive sentiments, there are topics, namely easy, all and accept. Furthermore, negative sentiment has 2 main issues such as access and problems. Finally, there are 3 neutral sentiments such as service, through and can. Future research is expected to use two or more methods or algorithms as a comparison in order to improve the shortcomings of the system.