Sentiment Analysis on Twitter using Neural Network: Indonesian Presidential Election 2019 Dataset

Ahmad Fathan Hidayatullah, Siwi Cahyaningtyas, Anisa Miladya Hakim

2021IOP Conference Series Materials Science and Engineering32 citationsDOI Open Access PDF

Abstract

Abstract Due to the utilization of Twitter by Indonesian politicians ahead of the Indonesian presidential election 2019, Indonesian people have given diverse responses and sentiments to the politicians. This study aims to classify sentiment on Indonesian presidential election 2019 tweet data by using neural network algorithms and obtain the best algorithm. In our study, we train our dataset using some variants of deep neural network algorithms, including Convolutional Neural Network (CNN), Long short-term memory (LSTM), CNN-LSTM, Gated Recurrent Unit (GRU) -LSTM and Bidirectional LSTM. Moreover, as a comparison with our deep learning model, we also train our dataset using other traditional machine learning algorithms, namely Support Vector Machine (SVM), Logistic Regression (LR) and Multinomial Naïve Bayes (MNB). Our experiments showed that Bidirectional LSTM achieved the best performance with the accuracy of 84.60%.

Topics & Concepts

IndonesianComputer scienceArtificial intelligenceConvolutional neural networkArtificial neural networkNaive Bayes classifierSupport vector machinePresidential electionDeep learningMachine learningMultinomial logistic regressionRecurrent neural networkSentiment analysisPoliticsLinguisticsPolitical sciencePhilosophyLawSentiment Analysis and Opinion MiningMultimedia Learning SystemsData Mining and Machine Learning Applications