Pengembangan Perangkat Lunak Klasifikasi Biner Sentimen Masyarakat Terhadap Vaksinasi Covid-19 Menggunakan Support Vector Machine

Abstract

This study aims to develop software for the binary classification of public sentiment towards the Covid-19 vaccination program which the Indonesian government began to organize on January 13, 2021. The data used comes from Twitter tweets with data crawling techniques. The tweets were classified binary through the training and testing stages which resulted in two classes of sentiment, namely positive and not positive. The feature extraction method used is Term Frequency-Inverse Document Frequency (TF-IDF), to weight the words in the form of a matrix based on the frequency of each word in the dataset. The method in the text classification process used is the Support Vector Machine (SVM). In the software development process, this research uses the incremental method. The results of the classification model training obtained model performance with an evaluation calculation of accuracy of 0.925, precision of 0.951, recall of 0.929, and f-measure of 0.94.