摘要
针对情感计算中传统的单模态情感分析通常存在分类准确率不高和不同语言环境间泛化能力较差的问题,提出一种双模态情感计算模型,以同时使用包含中英文两种语言、两种不同模态的情感数据。首先,利用多层感知机(MLP)网络和双向长短时记忆(BiLSTM)网络对数据进行特征提取;其次,基于MLP和自注意力机制分别对提取的特征进行特征融合,得到多模态分析模型;最后,使用该模型在构建的包含中英文两种语言数据的数据集上进行二分类情感计算预测。实验结果表明,所提模型相较于次优的BiLSTM模型,精度提高了1.22%;相较于单模态情感计算模型,精度提高了6.21%~14.00%。
Aiming at the traditional unimodal sentiment analysis in sentiment computing usually suffers from the problems of low classification accuracy and poor generalization ability between different language environments,a bimodal sentiment computation model was proposed to use the sentiment data containing both Chinese and English languages with two different modalities at the same time.Firstly,feature extraction was performed on the data using a MultiLayer Perception(MLP)network and a Bidirectional Long Short-Term Memory(BiLSTM)network.Sencondly,the extracted features were fused based on MLP and self-attention mechanism,respectively,to obtain a multimodal analysis model.Finally,the model was used to predict binary sentiment computation on a constructed dataset containing data in both Chinese and English languages.Experimental results show that:compared to the suboptimal BiLSTM model,the proposed model has the precision improved by 1.22%;compared to the unimodal sentiment computation models,the precision is improved by 6.21%-14.00%.
作者
吴俊洁
王佳阳
朱萍
肖强
WU Junjie;WANG Jiayang;ZHU Ping;XIAO Qiang(Zhejiang University Library,Zhejiang University,Hangzhou Zhejiang 310027,China;School of Electronics&Information,Hangzhou Dianzi University,Hangzhou Zhejiang 310018,China;Zhejiang Fangyuan Test Group Company Limited,Hangzhou Zhejiang 310018,China)
出处
《计算机应用》
CSCD
北大核心
2024年第S01期39-43,共5页
journal of Computer Applications
基金
国家自然科学基金资助项目(62001149)
国家重点研发计划项目(2020YFB1710600)
浙江省重点研发计划项目(2020C01110)。
关键词
情感计算
多语言泛化
多层感知机
自注意力机制
双模态
sentiment computing
multilingual generalization
MultiLayer Perception(MLP)
self-attention mechanism
bimodal
