基于话题标签的微博舆情热点挖掘系统设计  

Design of microblog public opinion hot spot mining system based on topic tags

作　　者：尚晓燕 SHANG Xiaoyan(Hohai University,Nanjing 210000,China)

机构地区：[1]河海大学,江苏南京210000

出　　处：《现代电子技术》2023年第2期70-74,共5页Modern Electronics Technique

摘　　要：为在微博舆情信息高速转发时段内实现对热点数据的精准挖掘,文中设计一种基于话题标签的微博舆情热点挖掘系统。首先确定框架体系连接形式,并根据热点信息定义结果建立话题标签模型。以此为基础,对热点信息散列表进行排重处理,完成对微博舆情热点数据的处理。最后,联合采集到的目标信息样本对论坛结构进行分割处理,完成挖掘系统设计。实验结果表明,在话题标签模型作用下,系统主机对于热点数据的挖掘准确率较高,可以在微博舆情信息高速转发时段内实现对热点数据的精准挖掘,满足实际应用需求;与基于机器阅读理解的挖掘系统、基于跨语言神经主题模型的挖掘系统相比,这种新型挖掘系统能够在微博舆情信息高速转发时段内,对热点数据进行精准挖掘,与设计初衷相符合。A microblog public opinion hot spot mining system based on topic tags is designed to achieve accurate mining of hot data during the high-speed forwarding period of microblog public opinion information. The connection form of the framework system is determined,and the topic tag model is established according to the definition results of hot information. On this basis,the hot information hash table is de duplicated,and the processing of hot data of microblog public opinion is completed. The forum structure is splited by combing with collected target information samples,so as to complete the design of the mining system. The experimental results show that under the action of the topic tag model,the system host has a high accuracy rate for hot data mining,which can achieve accurate mining of hot data in the high-speed forwarding period of microblog public opinion information to meet the actual application needs. In comparison with the mining system based on machine reading comprehension and the mining system based on cross language neural topic model, this new mining system can accurately mine hot data during the high-speed forwarding period of microblog public opinion information, which is consistent with the original design.

关 键 词：话题标签模型 微博舆情 高速转发 热点挖掘 目标信息 论坛结构 散列表 排重标准 

分 类 号：TN911-34[电子电信—通信与信息系统] TP393[电子电信—信息与通信工程]