基于CNN与GRU的文本情感分析研究

Research on Text Sentiment Analysis Based on CNN and GRU

利用CNN与GRU构建文本情感分析模型,预测影评的情感倾向,并对这些模型进行对比分析研究.结果表明,基于GRU的文本情感分析模型性能最好,在测试数据集上的准确率为86.77%,F1值为0.8723.

Text sentiment analysis,also known as opinion mining or opinion mining,is the process of analyzing,processing,and reasoning about subjective text with emotion.This paper uses CNN and GRU to construct text sentiment analysis models,predict the sentiment tendency of film reviews,and conduct comparative analysis and research on these models.The experimental results show that the GRU-based text sentiment analysis model has the best performance,with an accuracy rate of 86.77%and an F1 value of 0.8723 on the test set.