Sentiment Analysis of Tang Poetry by Fusing Text-graph Convolutional Neural Networks with Attention Mechanism
JIANG Tianqi
FANG Xianjin
REN Ping
Abstract：Aiming at the problem of insufficient semantic extraction and incomplete dataset in the current Tang poetry sentiment analysis task,we proposed a sentiment classification model integrating text graph convolutional neural networks with attention mechanism(AM-Text-GCN)by constructing a new Tang poetry sentiment classification dataset and further refining sentiment polarity.The model first combined bidirectional long short term memory(BiLSTM)with attention mechanism to capture contextual information and semantic features between Tang poetry sentences.Then,a two-layer text graph convolutional neural network incorporating dependency syntax analysis was used to aggregate the global features of Tang poetry in graph convolution operations,ultimately outputting the emotional polarity of Tang poetry.The results showed that the SF1 value of the proposed model reaches 79.83%,which is 5.46%higher compared to the SF1 value of text-graph convolutional networks,and effectively improved the accuracy of the sentiment analysis of Tang poems.This study is of great significance and has broad application prospects for exploring the role of poetic emotions in historical changes.
Keywords：sentiment analysisTang poetryText-GCNattention mechanismBiLSTM
Publication Date：2024-06-20
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：7( 205-211 )