Sentiment analysis of nuclear energy-related articles and their comments on a portal site in Rep. of Korea in 2010–2019

Soaring Energy Prices: Understanding Public Engagement on Twitter Using Sentiment Analysis and Topic Modeling With Transformers

“…The study found that ma-jority of the comments (71%) were neutral, followed by positive and negative comments that accounted for 23% and 6%, respectively. Opposite results have been reported in the research work conducted in [13] where the authors found that negative comments expressed by Korean people toward nuclear energy were larger than positive ones. Additionally, the study found that positive-tone articles were more present than negative ones.…”

“…• By dividing the output vectors by the embedding matrix, the vocabulary dimension is created. [22] 2023 Energy Twitter data Lexicon-based approach for sentiment analysis [23] 2023 Energy and other sectors Twitter data Twitter based investor sentiment index [16] 2022 Energy-related issues Twitter data (Word2Vec, GloVe and FastText), (DNN, LSTM, Bi-LSTM, CNN) [20] 2022 Renewable energy Twitter Lexicon-based Technique & CNN-LSTM [15] 2022 Nuclear energy Twitter data Decision Tree, Random Forest and LSTM [21] 2022 Renewable energy and GHG Twitter & Google Topic modelling through keyword similarity [18] 2021 Renewable energy Twitter data Transformers (RoBERTa) [13] 2021 Nuclear energy NAVER portal Lexicon-based approach [12] 2020 Nuclear energy Twitter data CNN, LSTM, and Bi-LSTM [10] 2020 Climate change & energy Twitter data Lexicon-based approach (EmoLex) [8] 2020 Energy Twitter data A lexicon-based and a fuzzy methods [19] 2019 Renewable Energy Twitter data SVM, KNN, Naïve Bayes, AdaBoost, Bagging [14] 2019 Green buildings Sina Weibo Lexicon-based approach [7] 2018 Energy Twitter data Two sentiment lexica [11] 2018 Nuclear Energy Twitter data Attentive Deep Neural Network FIGURE 1: BERT using Masked Language Modeling [24] • Use sof tmax to determine the likelihood of each word in the lexicon.…”