摘要
网络中充斥着许多带有强烈情感的评论和信息,对这些信息的分析成为了解人们意见的主要途径。情感分析是自然语言处理(natural language processing,NLP)的一个方向,用来判断文本描述的情绪类型。通过建立用户自身的情感标志模型来识别语句的情感特征,并将设计好的模型利用深度学习框架进行情感分析,最终得到各模型的准确率。利用TensorFlow深度学习框架,对公开数据集分别采用CNN、LSTM模型进行情感分析和比较研究。验结果表明,LSTM模型在实验中表现较佳并可为相关情感分析模型的优化提供一定的意义。
In the Internet there are full of comments and information with strong emotions,and the analysis of such information becomes the main way to know people's opinions.Emotion analysis is a direction of natural language processing(NLP),which is used to judge the types of emotions described in text.The emotional feature of the sentence is identified by building the user's own emotional model,and the designed model is used for emotional analysis with the deep learning framework.Finally,the accuracy of each model is obtained.Using the TensorFlow deep learning framework,the open data sets were analyzed and compared using CNN and LSTM models.The experimental results show that the LSTM model performs better in the experiment and can provide some significance for the optimization of related emotion analysis model.
作者
杨丹
张梦
朱毅
YANG Dan;ZHANG Meng;ZHU Yi(Dalian University of Foreign Languages,Dalian 116041,China)
出处
《电脑知识与技术》
2019年第7X期188-190,共3页
Computer Knowledge and Technology
基金
大连外国语大学学生创新创业训练计划项目(201810172091)的资助
关键词
情感分析
卷积神经网络
长短期记忆神经网络
sentiment analysis
convolutional neural network
long short-term memory network
