Implicit Sentiment Analysis of Short Stories Based on GT-BiLSTM

This paper proposes an implicit sentiment analysis method for short stories based on the GT -BiLSTM model. Sentiment analysis has been widely used in natural language processing (NLP), especially in social media and literature. However, the emotions in short stories are usually implicit, and traditional sentiment analysis methods are difficult to effectively capture these complex emotional expressions. In order to solve this problem, this paper proposes a bi-directional short-term and long-term memory network (GT-BiLSTM) model that combines the graph attention mechanism and the emotion triad. The model improves the accuracy and depth of sentiment analysis by effectively capturing the emotional relationships and contextual information in the text. The experimental results show that GT-BiLSTM has good potential for short story sentiment analysis, which can provide a new technical path for literary sentiment analysis and open up a new direction for the application of sentiment analysis in more complex text types.