부산대학교 도서관

The proliferation of massive amounts of raw textual data, particularly through social media, has made sentiment analysis more challenging in the age of big data. Due to the significance of consumer feedback, sentiment analysis has become increasingly important in a variety of industries, including retail, dining, and travel. This study investigates the use of deep learning techniques for sentiment analysis, including bidirectional long short-term memory (Bi-LSTM), and gated recurrent unit (GRU). The study also investigates how well word embedding techniques like BERT and RoBERTa capture the subtleties of sentiment, giving academics and business experts insights into how to use deep learning techniques for sentiment analysis in practical contexts. To evaluate the performance of these techniques, we consider two datasets: Twitter and IMDB movie reviews. The sentiment analysis process aims to classify the sentiment of the texts as positive or negative, while also determining the sarcasm score associated with each sentiment.By contrasting the outcomes of different deep learning architectures and word embedding techniques on diverse datasets, this research sheds light on their effectiveness in sentiment analysis tasks. The findings contribute to a better understanding of the strengths and limitations of each approach, offering insights for researchers and practitioners seeking to leverage deep learning techniques for sentiment analysis in real-world applications.