摘要
为提高舆情风险预测的准确性和效率,提出一种基于注意力机制与双向长短期记忆(BiLSTM)网络相结合的舆情感知模型。该方法通过结合BiLSTM的双向建模能力与注意力机制的特征聚焦能力,精准捕捉舆论数据中的情感波动和上下文语义特征,从而提升舆情风险的预测精度。同时以“高考冒名顶替”事件为样本,展开网络舆论数据分析。通过与ELM、随机森林、决策树、LSTM、BiGRU和BiLSTM等多种主流算法进行对比实验,验证所提模型的有效性和优越性。在算法设计中,注意力机制的引入有效地提升了模型在长文本情感分类中的表现,能够精确捕捉情感变化的关键节点。实验结果表明,所提出的预测模型能够有效地识别出舆情风险,准确率达到94.87%,相比于表现最优的BiGRU算法提高了约5.75%。
In order to improve the accuracy and efficiency of public opinion risk prediction,a public opinion perception model based on the combination of attention mechanism and bidirectional long short-term memory(BiLSTM)network is proposed.This method can be used to accurately capture emotional fluctuations and contextual semantic features in public opinion data by combining the bidirectional modeling ability of BiLSTM with the feature focusing ability of attention mechanism,so as to improve the prediction accuracy of public opinion risks.By taking the"college entrance examination impersonation"incident as a sample,network public opinion data was analyzed.The effectiveness and superiority of the proposed model are verified by means of comparative experiments with various mainstream algorithms such as ELM,random forest,decision tree,LSTM,BiGRU and BiLSTM.In algorithm design,the introduction of attention mechanism can effectively improve the performance of the model in long text emotion classification,and can accurately capture key nodes of emotional changes.The experimental results show that the proposed prediction model can effectively identify public opinion risks,with an accuracy of 94.87%,which is about 5.75%higher than the best-performing BiGRU algorithm.
作者
王敏
汪旭
WANG Min;WANG Xu(Nanchang University,Nanchang 330031,China;Nanchang Hangkong University,Nanchang 330063,China)
出处
《现代电子技术》
北大核心
2025年第12期115-119,共5页
Modern Electronics Technique
基金
2023年江西省社会科学基金项目(23XW09)。
关键词
舆情风险预测
情感识别
自然语言处理
双向长短期记忆网络
注意力机制
文本分类
public opinion risk identification
emotional identification
natural language processing
bidirectional long shortterm memory network
attention mechanism
text classification
作者简介
王敏(1992-),女,江西南昌人,博士研究生,研究方向为新闻信息传播与分析技术;汪旭(1994-),男,江西鹰潭人,博士研究生,讲师,研究方向为传播与分析技术。
