Sentiment Analysis Public Twitter on 2024 Election using the Long Short Term Memory Model

Abstract

As a Democratic State in 2024 is a political year for Indonesia because there will be simultaneous elections starting from officials from the regions to the president, therefore it causes a lot of diverse reactions on social media, especially Twitter which can give rise to various kinds of opinions, not only positive but also positive opinions. also negative ones so that it becomes an interesting thing to study, the tendency of Twitter users to post content can be known through sentiment analysis, so the purpose of this study is to find out the sentiment of twitter users towards the 2024 election is negative or positive, the steps in this study begin with data collection tweets, preprocessing text mining and then entering lexicon based feature labels, building a model with Long Short Term Memory (LSTM) and evaluating the model with the confusion matrix tool used in this study using google collaboratory, sentiment analysis is carried out with gan 858 data from the preprocessing process with a comparison of 80% of train data and 20% of test data resulted in an accuracy of 78% with positive sentiment amounting to 52.2%, sentiment that was included in the positive mostly contained support for the success of the election, as well as information related to the election, and 37% negative sentiment which is included in the negative sentiment in this study include tweets with dirty words, misguiding opinions, and making fun of political and neutral parties by 10.8%. Keywords: 2024 Election, Sentiment Analysis, twitter, LSTM