摘要
基于机器学习的情感分类方法已经取得了较大进展,但在大量情感分类方法中,往往都是结合词嵌入和传统的机器学习方法,缺乏对文本主题以及时序关系等因素的有效利用。针对上述问题,提出了一种基于主题流与深度学习的情感分类算法,通过分析文本的主题分布,并引入时序关系,在此基础上利用适合的长短记忆神经网络的深度学习方法进行情感分类。实验证明,基于主题流与深度学习的情感分类算法性能较好。
At present,sentiment classification method based on machine learning has made great progress,but among the a large number of sentiment classification methods,word combination and traditional machine learning methods are often used,and there is a lack of effective use of such factors as text topics and sequence relationships in a large number of sentiment classification methods.In order to solve the problems,this paper presents a sentiment classification algorithm based on topic flow and deep learning,it analyzes the topic distribution of the text and introduces the sequence relationship and uses deep learning methods such as long short-term memory neural networks to classify the sentiment.Experiments show that the sentiment classification algorithm based on topic stream and deep learning proposed in this paper has better performance.
作者
刘纳
王新
LIU Na 1,WANG Xin 2(1.College of Computer Science and Engineering,Shandong University of Science and Technology,Tsingtao266590; 2.Jinan Software Research Institute,China United Network Communications Limited,Jinan 250199,Chin)
出处
《软件导刊》
2018年第8期28-30,34,共4页
Software Guide
