基于微博数据的云南省地理情感及主题特征研究

Study on Geographical Emotion and Topic Characteristics of Yunnan Province Based on Weibo Data

基于情感词典的情感定量化计算及基于BTM模型(双语词话模型)的主题建模方法,分析了云南省2021年3月~5月的微博情感分布、主题的分布及二者之间的联系。结果发现:云南省129个区县的微博情感分布具有明显的冷热点(高/低值聚类)效应,且存在两个冷点聚集区和两个热点聚集区;微博情感热点区域的共性是关注旅游,并且有较多的正面情绪或较少的负面情绪;微博情感冷点区域的共性是更加关注离乡,不同的是昭通市境内的微博情感冷点区对疾病有更多的关注,且负面情绪分布较多。

Based on the quantitative calculation of emotion in emotion dictionary and the topic modeling method based on BTM model(bilingual dictionary model),this paper analyzes the emotional distribution,topic distribution and the relationship between them in Weibo from March to May,2021.The results show that the emotional distribution of Weibo in 129 districts and counties of Yunnan Province has obvious cold hot spot(high/low value clustering)effect,and there are two cold spot clusters and two hot spot clusters;The commonness of emotional hotspots in Weibo is that they pay attention to tourism and have more positive emotions or less negative emotions;The commonness of emotional cold spots in Weibo is that they pay more attention to leaving their hometown,but the difference is that the emotional cold spots in Weibo in Zhaotong pay more attention to diseases and have more negative emotions.