摘要
利用大连理工大学情感词汇本体库DUTIR对福建地区居民关于降雨的微博进行细粒度情感分析,将情感倾向分为“乐、好、怒、哀、惧、恶、惊”7类,并得到相应情感倾向度。情感分析的准确率为92.75%,平均综合评价指标为68.28%,结合关键词挖掘技术进一步分析,结果发现:(1)福建地区居民对降雨的情感倾向主要为“好”,总体呈正面情感;(2)福建地区居民对降雨的情感是多样且稳定的;(3)相关天气现象、所处地点、环境感受、社会活动影响构成了影响福建地区居民对降雨情感的因素。
Using DUTIR to conduct fine-grained sentiment analysis of microblogs about rainfall of residents in Fujian area,the sentiment tendencies are divided into seven categories,including happy,good,anger,sadness,fear,disgust,and surprise,and the corresponding sentiment tendencies are obtained.The accuracy rate of sentiment analysis is 92.75%,and the average comprehensive evaluation index is 68.28%.Combined with keyword mining technology for further analysis,it is found that,(1)Residents in Fujian aera have a“good”emotional tendency toward rainfall,and generally present positive emotions.(2)Emotions toward rainfall are diverse and stable.(3)Relevant weather phenomena,location,environmental feelings,and social activity influences constitute the factors that affect the feelings of the residents in Fujian for rainfall.
作者
余安安
吴雪菲
李栋
任雍
刘光普
YU An'an;WU Xuefei;LI Dong;REN Yong;LIU Guangpu(Fujian Key Laboratory of Severe Weather,Fuzhou 350008,China;Fujian Atmospheric Detection Technology Support Center,Fuzhou 350008,China)
出处
《亚热带资源与环境学报》
2022年第4期29-36,共8页
Journal of Subtropical Resources and Environment
基金
福建省气象局青年科技专项“基于神经网络的毫米波测云雷达云状识别研究”(2022Q06)。
关键词
细粒度情感分析
福建地区
微博
降雨
关键词挖掘
fine-grained sentiment analysis
Fujian area
Weibo
rainfall
keyword mining
