摘要
本文重点是对网络用户中的网络舆情倾向性进行分析,构建一种准确率更高的模型来实现情感倾向判断。网络舆情倾向性分析主要是通过机器学习的方法和制定部分规则的方法,从而能够更加准确地判断出句子的情感倾向性,进而能在实际生活中进行有针对性地积极引导调控。基于此,提出一个设想,对于机器学习的方法之间的结合与本身的改进,再加入规则的制定,能够提高机器对情感句子的分析准确率。对于模型应用之前的特征提取的方法进行比较选取,通过特定的特征选取方式使得构造向量空间更加准确,再通过比较不同分类算法进行比较选取。
This paper focuses on the analysis for Internet public opinion orientation of online users to conduct a more accurate modeling of emotional tendency classification. Through the machine learning method and the method of regulating partial rules, the Internet public opinion orientation analysis can predict the emotional tendency of sentences accurately, thus actively benefit the real life. As a result, it is deduced that through the serf-improvement, the machine learning method with the regulating of rules can improve the accuracy for analysis of emotional sentence. The paper selects the feature extracting method before applying the model which accurately constructs the vector space, and also realizes the selection among different classification algorithms,
出处
《智能计算机与应用》
2017年第2期127-130,共4页
Intelligent Computer and Applications
关键词
网络舆情
机器学习
特征提取
支持向量机
public opinion
machine learning
feature extraction
Supporting Vector Machine
作者简介
汪淳（1992-），男，硕士研究生，主要研究方向：自然语言处理。
