基于情感特征和AM的新媒体精准传播检测技术研究  

Research on accurate communication detection technology of new media based on emotional characteristics and AM

作　　者：韩鹏[1] 许萌 HAN Peng;XU Meng(Xianyang Vocational Technical College,Xianyang Shaanxi 712000,China)

机构地区：[1]咸阳职业技术学院,陕西咸阳712000

出　　处：《自动化与仪器仪表》2025年第3期211-214,219,共5页Automation & Instrumentation

基　　金：咸阳职业技术学院科研基金资助项目(2024 KJB02)。

摘　　要：为精准检测新媒体文本传播中的谣言问题,提出一种基于情感特征和AM相结合的谣言检测方法。首先,采用BERT模型中的两层Transformer编码器对谣言语义特征进行提取;然后通过BiLSTM-AM进行文本情感特征提取;最后将语义特征向量与情感特征向量进行拼接融合,并将其输入至LSTM模型中进行新媒体文本谣言分类,实现新媒体文本谣言的精准检测。实验结果表明,本模型的新媒体文本谣言检测准确率、召回率和F1分数分别取值为97.42%、96.58%和98.03%,明显高于BERT-GNNs模型、MFCC-MLA模型和BiLSTM-CNN-ECA模型,且本模型的检测时长仅为7.49 s,均低于另外三种模型。由此说明,本模型能够提升新媒体传播文本谣言检测准确率和效率,检测效果显著提升。In order to accurately detect the rumors in the dissemination of new media text,a rumor detection method based on emotional characteristics and AM is proposed.First,two-layer Transformer encoder in BERT model;then through BiLSTM-AM;finally,semantic feature vector and input into the LSTM model for new media text rumor classification,so as to realize accurate detection of new media text rumors.The experimental results show that the rumor detection accuracy,recall rate and F1 score of this model are 97.42%,96.58% and 98.03%,respectively,which are significantly higher than BERT-GNNs model,MFCC-MLA model and BiLSTM-N-N-ECA model,and the detection time of this model is only 7.49 s,which is lower than the other three models.This shows that this model can improve the accuracy and efficiency of new media propagation text rumor detection,and significantly improve the detection effect.

关 键 词：情感特征 注意力机制 BiGRU 新媒体 传播检测 

分 类 号：TP392[自动化与计算机技术—计算机应用技术]