融合注意力机制和Bi-LSTM的情感分析  

Emotion Analysis Based on Attention Mechanism and Bi-LSTM

作　　者：邓翔 Deng Xiang(CETC,Beijing,China)

机构地区：[1]电科院,北京

出　　处：《科学技术创新》2025年第9期93-96,共4页Scientific and Technological Innovation

摘　　要：针对微博评论长短不一、用语复杂和表达隐晦等现象,本文提出了融合多头注意力机制和Bi-LSTM深度学习模型的评论数据情感分析方法,通过对23万条微博评论数据进行实验验证,能够有效识别评论中的情感倾向,对于复杂的文字描述也有一定的鲁棒性。In response to the phenomenon of inconsistent length,complex language,and ambiguous expression in Weibo comments,this paper proposes a sentiment analysis method for comment data that integrates Multi-Head Attention mechanism and Bi-LSTM deep learning model.Through experimental verification of 230000 Weibo comment data,it can effectively identify emotional tendencies in comments and has certain robustness to complex textual descriptions.

关 键 词：情感分析 Bi-LSTM 注意力机制 微博评论 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]