| 融合语义与情感分析的区块链产业新闻监测研究 |
| |
| 引用本文： | 吴俊,邵丹睿,姜尚杨帆.融合语义与情感分析的区块链产业新闻监测研究[J].现代情报,2021,40(11):22-33. |
| |
| 作者姓名： | 吴俊 邵丹睿 姜尚杨帆 |
| |
| 作者单位： | 北京邮电大学经济管理学院, 北京 100876 |
| |
| 基金项目： | 国家社会科学基金"推动新一代信息技术与制造业深度融合研究——基于新时代和新工业革命的视角"（项目编号：18VSJ054）；国家重点研发计划项目"基于模式创新的科技咨询服务平台研发与应用示范"（项目编号：2018YFB1403600）；北京市社会科学基金规划项目"基于大数据的北京市共享单车产业监测与发展趋势研究"（项目编号：17YJB018）。 |
| |
| 摘 要： | 目的/意义] 前沿技术孵育的新兴产业发展演进快，但因统计数据迟滞，产业监测难而备受研究者关注。方法/过程] 以2014-2019年36氪网站互联网区块链新闻为数据样本，提出纳入协变量的结构化主题模型（STM）与深度学习情感分析技术结合的新兴产业新闻文本监测方法，通过监测媒体报道的产业新闻热点强度变化，文本情感倾向对新闻热点强度的时序影响，发现并跟踪新兴产业热点及趋势。结果/结论] 2014-2019年，69%的区块链新闻主题聚焦于区块链的产业应用和比特币等数字代币的发行与交易。文本的语义和情感分析显示，2017年以来，中国的区块链产业发展存在一定的媒体炒作特征，但媒体对各类数字代币发行与交易由褒转贬的情感倾向变化可以对区块链隐含风险起到预警作用。创新/价值] 提出的产业新闻文本监测方法具有准实时性，能与传统的事后统计指标监测方法互为补充。
|
| 关 键 词： | 区块链 产业新闻 结构化主题模型 文本情感分析 深度学习 |
Identifying Development Focus and Trend of Blockchain Industry Through News Text Mining: A Topic Modeling and Sentiment Analysis Investigation |
| |
| Authors: | Wu Jun Shao Danrui Jiang Shangyangfan |
| |
| Institution: | School of Economics and Management, Beijing University of Posts and Telecommunications, Beijing 100876, China |
| |
| Abstract: | Purpose/Significance] How to identify the development focus and trend of emerging technology,like blockchain,and its industry have attracted more and more attention from academics and practitioners.Method/Process] 3983 blockchain news collected from China famous TMT media-36kr.com was analyzed using structural topic modeling and deep learning enhanced sentiment analysis.Results/Conclusions] This paper revealed that:(1)blockchain industry development in China during the period of 2014 to 2019 can be reflected by 7 key topics which focused on the domain applications of blockchain technology,initial offering and transaction of bitcoin and digital currency.(2)Although positive sentiment tendency dominated during the research period of 2014 to December 2019,the sentiment of news related to the issuance and transaction of digital currency has clearly shifted from positive to negative since January 2017,implying concerns about speculation in the encrypted digital currency market.The contribution of this paper is reflected not only by proposing a new approach to identify the development focus and trend of technology frontiers but also validating its application in the blockchain industry monitoring. |
| |
| Keywords: | blockchain industry news structural topic modeling sentiment analysis deep learning |
|
| 点击此处可从《现代情报》浏览原始摘要信息 |
| 点击此处可从《现代情报》下载免费的PDF全文 |
