Sentimen Analisis Indihome di X (Twitter) dengan Metode Fasttext dan Support Vector Machine

Abstract

Indonesia Digital Home or commonly abbreviated as Indihome is one of the internet, home telephone and television service products that implement fiber optic cable from PT Telekomunikasi Indonesia. IndiHome is the most widely used internet service provider in Indonesia due to its wide signal coverage and attractive package prices. The high number of users is directly proportional to the community's need to access various information as the internet has become a primary need. Despite all that, IndiHome still cannot be separated from service interruptions. Interference with IndiHome consists of loss interference and router interference. Every time a disturbance occurs, users will report to Indihome customer service available via telephone 188, Whatsapp application, MyIndihome application and email. Various ways users complain and criticize IndiHome services, one of which is using sosial media X (twitter). The opinions contained in X (tweet) are quite varied, this is because X (twitter) users have different backgrounds and come from different groups of society. So it is necessary to analyze the sentiment of Indihome users on X (Twitter) using the FastText and Support Vector Machine (SVM) methods. The classification process starts from dataset input, data pre-processing, data labeling, split data, word embedding, classification and evaluation. The results of this study show that evaluation using confusion matrix without using FastText (TF-IDF) can produce an accuracy value of 56.11% higher than using FastText which is 55.78%.