부산대학교 도서관

Sentiment analysis is essential for comprehending the underlying emotions and opinions conveyed in written information. For this purpose, conventional techniques like Convolutional neural networks (CNN) and recurrent neural networks (RNN) have been widely used. This paper aims to investigate novel techniques to sentiment analysis by merging Bidirectional Long Short-Term Memory (BiLSTM) networks with Conditional Random Fields (CRF) to find explicit characteristics. The study assesses the effectiveness of the BiLSTM-CRF model by comparing its performance to that of commonly used models such as CNN and RNN. The main goal is to examine the effectiveness of the BiLSTM-CRF model in detecting explicit aspects of sentiment in textual data and determine if it performs better than established methods. This study involved conducting multiple experiments on benchmark datasets and evaluating the models using various metrics such as accuracy, precision, and Fl score. Initial results indicate the BiLSTM-CRF model shows promising performance of up to 85% accuracy, suggesting its ability to improve sentiment analysis tasks. This study enhances sentiment analysis techniques by investigating novel methodologies and revealing the efficacy of Bidirectional Long Short-Term Memory (BiLSTM) networks with Conditional Random Fields (CRF) for explicit aspect identification.