Sentiment Analysis On Twitter Posts About The Russia and Ukraine War With Long Short-Term Memory

Abstract

Sentiment analysis is one method for evaluating public opinion from the received text. In this study, we evaluate the performance of the LSTM model with Sastrawi in sentiment analysis in Indonesian using a Twitter dataset totaling 2537 data collected regarding the Russo-Ukrainian war. The purpose of this study is to determine the reliability of the LSTM model with Sastrawi in sentiment analysis in Indonesian and to evaluate the performance of the model with the collected Twitter dataset regarding the Russian-Ukrainian war. The method used in this study is data pre-processing, training and validation of the LSTM model with Literature, and model evaluation using the metrics of accuracy, precision, recall, and F1 score. In the dataset collected in this study, positive, neutral and negative sentiments were 54.7%, 35% and 10.2%. The results obtained from this study indicate that the LSTM model with Literature can provide good results in sentiment analysis with a prediction accuracy of 82%. The implication of the results of this study is that the LSTM model with Sastrawi can be used for sentiment analysis on Twitter and further research needs to be carried out with a wider and more diverse dataset, especially to produce even better accuracy.