基于GRU的新媒体语义分析方法研究
Research on new media semantic analysis method based on GRU
刘恋1
作者信息
- 1. 奥克兰大学人文学院,新西兰奥克兰1010
- 折叠
摘要
传统情感分析很大程度上依赖于情感词典的质量和特征提取,这极大限制了媒体情感语义的正确表达.直至有了深度学习,才使新媒体语义分析得以进步.文中构建了基于GRU的新媒体语义快速分析模型,且针对相关模型进行数据训练,并将其应用于实际案列.经与传统模型对比,得出以下结论:较SVM、LSTM模型而言,GRU模型分析能力更强,分类准确率更高,高达0.909;GRU模型情感极性判断能力更佳,但在中性词的判断上略有欠缺;GRU模型预测能力更强.
Abstract
Traditional sentiment analysis largely relies on the quality of sentiment dictionaries and feature extraction,which greatly limits the correct expression of media sentiment semantics.It is not until deep learning that semantic analysis of new media is advanced.The article constructs a new media semantic fast analysis model based on GRU,data training is conducted on relevant models and applied to practical cases.After comparison with traditional models,the following conclusions are drawn:Compared with SVM and LSTM models,the GRU model has stronger analytical ability and higher classification accuracy,up to 0.909;The GRU model has better emotional polarity judgment ability,but slightly lacks in neutral word judgment;The GRU model has stronger predictive ability.
关键词
新媒体/深度学习/GRU/语义分析/情感Key words
new media/deep learning/GRU/semantic analysis/emotion引用本文复制引用
出版年
2024
NETL
NSTL
万方数据