基于word2vec和双向LSTM的情感分类深度模型  

Sentiment classification depth model based on word2vec and bi-directional LSTM

作　　者：黄贤英[1] 刘广峰 刘小洋[1,2] 阳安志 Huang Xianying;Liu Guangfeng;Liu Xiaoyang;Yang Anzhi(College of Computer Science&Engineering,Chongqing University of Technology,Chongqing 400054,China;College of Engineering,University of Alabama,Tuscaloosa Alabama 35401,USA)

机构地区：[1]重庆理工大学计算机科学与工程学院,重庆400054 [2]College of Engineering,University of Alabama,Tuscaloosa,Alabama 35401,USA

出　　处：《计算机应用研究》2019年第12期3583-3587,3596,共6页Application Research of Computers

基　　金：重庆理工大学研究生创新项目(ycx2018245);重庆市教育委员会人文社会科学研究项目(17SKG144);国家教育部人文社科青年基金资助项目(16YJC860010);国家社科基金资助项目(17XXW004);2018年重庆市科委技术创新与应用示范项目(cstc2018jscx-msybX0049)

摘　　要：针对社交网络文本传统情感分类模型存在先验知识依赖以及语义理解不足的问题,提出一种基于word2vec和双向长短时记忆循环神经网络的情感分类模型--WEEF-BILSTM。采用基于CBOW (continuous bag-of-words)方式的word2vec模型针对语料训练词向量,减小词向量间的稀疏度,通过双向LSTM神经网络获取更为完整的文本上下文信息从而提取出深度词向量特征,继而使用one-versus-one SVM对其进行情感分类。实验结果表明,提出的WEEF-BILSTM模型较其他模型分类效果更好,能达到更优的准确率和F值。Aiming at the problem of prior knowledge and lack of semantic understanding in the traditional sentiment classification model of social network text,this paper proposed a sentiment classification model called WEEF-BILSTM based on word2 vec and bi-directional long short-term memory recurrent neural network. It used the CBOW-based word2 vec model for corpus to train word vectors,reduced the sparsity between word vectors,and obtained more complete text context information through the bi-directional LSTM neural network to extract the depth word vector features,and then used one-versus-one SVM to classifiy emotions. The experimental results show that the proposed WEEF-BILSTM model has better classification results than other models,and can achieve better accuracy and F value.

关 键 词：文本分类 情感分析 双向长短时记忆循环神经网络 词向量 社交网络 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]