基于SVM算法的航天微博情感分析  

Sentiment Analysis of Aerospace Microblog Using SVM

作　　者：徐建忠 朱俊[2] 赵瑞[3] 张亮 何亮 李娇娇 Xu Jianzhong;Zhu Jim;Zhao Rui;Zhang Liang;He Liang;Li Jiaojiao(Hangzhou Shi ping Information & Technology Co., Ltd,Hangzhou 310012;Xichang Satellite Launch Center, Wenchang, Hainan 571300;Acdemy of Armored Forces Engineering, Beijing 100072)

机构地区：[1]杭州世平信息科技有限公司,杭州310012 [2]西昌卫星发射中心,海南文昌571300 [3]装甲兵工程学院,北京100072

出　　处：《信息安全研究》2017年第12期1129-1133,共5页Journal of Information Security Research

摘　　要：微博作为人们获取和传播信息的重要平台,是分析掌握大众舆情的重要数据源.为了掌握航天事件对舆论的影响,对航天相关微博数据进行情感分析,通过人工构建训练数据集、特征设计和SVM模型训练,建立了判断航天相关微博情感正极性和负极性的方法.实验结果表明:该方法有着较高的准确率和召回率,验证了所提方法在航天相关微博情感分析判断方面的有效性.Microblog is an important platform,from which people can obtain and transmit information,and also is important data source to analyze and grasp the public opinion.In order to grasp aerospace event's influence on public opinion,this paper carried out a study on aerospace microblog data sentiment analysis,via manual training data construction,feature designing and support vector machine(SVM)model training,we propose a method to judgment the aerospace microblogging is positive or negative.Experiment result show that,the method discussed in this paper get high precision and recall value,which verify the effectiveness of proposed method in aerospace microblogging sentiment analysis judgment.

关 键 词：情感分析 航天 微博 特征设计 SVM算法 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]