您的位置:首页  > 论文页面

基于BiGRU-CNN的中文评论文本情感分析

发表时间:2019-06-28  浏览量:2914  下载量:594
全部作者: 张苗,张征
作者单位: 华中科技大学人工智能与自动化学院
摘 要: 目前,常用的深度学习模型是基于卷积神经网络(convolutional neural network,CNN)和循环神经网络(recurrent neural network,RNN)的模型,而RNN中包括长短期记忆网络(long-short term memory,LSTM)和门控循环单元(gated recurrent unit,GRU)两种变体模型。为提高评论文本情感分析的准确率,本文将RNN与CNN结合,提出一种融合双向门控循环单元(bidirection gated recurrent unit,BiGRU)和CNN的BiGRU-CNN文本情感分析模型。该模型在商品评论和电影评论数据集上的准确率分别达到了93.27%和90.76%. 相比CNN、LSTM和GRU等基本模型,BiGRU-CNN模型提高了文本情感分析的准确率,并且模型训练时间适中。
关 键 词: 人工智能;情感分析;深度学习;门控循环单元;卷积神经网络
Title: Sentiment analysis of Chinese comment text based on BiGRU-CNN
Author: ZHANG Miao, ZHANG Zheng
Organization: School of Artificial Intelligence and Automation, Huazhong University of Science and Technology
Abstract: At present, the commonly used deep learning model is based on convolutional neural network (CNN) and recurrent neural network (RNN), and RNN includes two structures: long-short term memory (LSTM) and gated recurrent unit (GRU). In order to improve the accuracy of sentiment analysis of comment text, RNN is combined with CNN, and a bidirection gated recurrent unit (BiGRU)-CNN text sentiment analysis model which integrates BiGRU and CNN is proposed in this paper. The accuracy of this model on the data sets of product reviews and film reviews reached 93.27% and 90.76% respectively. Compared with the basic models of CNN, LSTM and GRU, BiGRU-CNN model improves the accuracy of sentiment analysis of text and the training time of the model is moderate.
Key words: artificial intelligence; sentiment analysis; deep learning; gated recurrent unit; convolutional neural network
发表期数: 2019年6月第3期
引用格式: 张苗,张征. 基于BiGRU-CNN的中文评论文本情感分析[J]. 中国科技论文在线精品论文,2019,12(3):363-370.
 
6 评论数 0
暂无评论
友情链接