基于Elmo和注意力机制的双通道文本分类模型  

Dual Channel Text Classification Model Based on Elmo and Attention Mechanism

作　　者：陈小莹[1] 艾金勇[2,3] CHEN Xiao-ying;AI Jin-yong(School of Information Engineering,Xizang MinzuUniversity,Xianyang Shaanxi 712082,China;Library of Xizang Minzu University,Xianyang Shaanxi 712082,China;Key Laboratory of Optical Information Processing and Visualization Technology of Tibet Autonomous Region,Xianyang Shaanxi 712082,China)

机构地区：[1]西藏民族大学信息工程学院,陕西咸阳712082 [2]西藏民族大学图书馆,陕西咸阳712082 [3]西藏自治区光信息处理与可视化技术重点实验室,陕西咸阳712082

出　　处：《计算机仿真》2024年第10期507-512,523,共7页Computer Simulation

基　　金：西藏民族大学校内项目(22MDY022)。

摘　　要：针对中文文本分类过程中文本特征提取不全面、语义表征不准确的问题,提出一种基于改进Elmo模型、带有注意力机制的卷积神经网络与门控循环网络相结合的双通道文本分类模型。模型首先将静态词向量输入Elmo模型生成动态词向量对文本进行表示;然后利用双通道结构构建加入注意力机制的卷积神经网络和双向门控循环网络分别提取文本内部特征和全局语义信息;最后,将双通道特征向量融合处理后通过分类器完成文本分类。依托THUCNews数据集进行模型的仿真,所提模型分类准确率和召回率分别为90.21%、90.45%,实验结果表明,与其它分类模型相比,所提模型具有更好的分类性能。A dual channel text classification model based on an improved Elmo model,a convolutional neural network with attention mechanism,and a gated recurrent network is proposed to address the problems of incomplete feature extraction and inaccurate semantic representation in the process of Chinese text classification.In this model,static word vector is input into Elmo model to generate dynamic word vector to represent text.Then,the dual channel structure is used to construct the convolutional neural network and the bi-directional gated recurrent unit network that join the attention mechanism to extract the internal features and global semantic information of the text respectively.Finally,the dual channel feature vectors are fused to complete the text classification through the classifier.Based on the THUCNews data set,the simulation experiment of the model is carried out.The classification accuracy and recall rate of the proposed model are 90.21%and 90.45%respectively.The experimental results show that the proposed model has better classification performance than other classification models.

关 键 词：文本分类 特征融合 注意力机制 双通道 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]