摘要
突变词具有前瞻性和情报意义,研究主题突变可预测学科领域的研究前沿和研究热点。本文构建突变词检测多维指标体系,包括无序性、增长性、突变度三个突变特征指标,以及知识融合度指标和影响力度指标。基于无序性和增长性两个维度,运用K-means聚类出突现型关键词(突现词)、强突型关键词(强突词)、弱突型关键词(弱突词)三类突变词。结合各类突变词的突变度、知识融合度和影响力度,识别不同发展形态的突变词。结果表明,高突变度的突现词在初期能获得更多的关注,形成更大的影响力;知识融合度高的突变词具有较高的交叉融合广度和强度,未来更可能发展成为研究热点;影响力度高的突变词具有广泛的关注和一定的研究基础,未来有更高概率成为研究前沿。
Since burst terms are forward-looking and informative,burst term detection(BTD)helps to predict research fronts and hotspots in certain subject areas.In this study,we build a multi-indicator system for BTD,including burst indicators(“random,”“growth,”and“burst”),knowledge fusion indicators,and influence indicators.Based on“random”and“growth”,three categories of burst terms are clustered by K-means,namely,emergence terms,strong burst terms,and weak burst terms.Combining the burst indicators,knowledge fusion indicators,and influence indicators,the burst terms with different developmental statuses were identified.The results show that emergence terms with a high burst can gain more attention and have more influence at the initial stage.Burst terms with a high degree of knowledge fusion indicate that the breadth and intensity of fusion are higher,and they are more likely to develop into research hotspots in the future.Finally,burst terms with high influence indicates that they receive wide attention and have a certain research base and have a higher probability to become a research frontier in the future.
作者
彭国超
孔泳欣
王玉文
Peng Guochao;Kong Yongxin;Wang Yuwen(School of Information Management,Sun Yat-sen University,Guangzhou 510006)
出处
《情报学报》
CSSCI
CSCD
北大核心
2022年第6期584-593,共10页
Journal of the China Society for Scientific and Technical Information
基金
国家自然科学基金项目“智慧城市App用户使用行为分化机理研究”(71974215)
中山大学高校基本科研业务费青年教师重点培育项目“科学计量学视角下智慧城市研究热点、发展趋势及跨学科交叉融合分析”(20wkzd17)。
关键词
主题突变
突变检测
多指标融合
聚类分析
burst term
burst term detection
multi-indicators
cluster analysis
