基于LDA模型与ATM模型的学者影响力评价研究——以我国核物理学科为例

Scholar Influence Evaluation Research Based on LDA Model and ATM Model——A Case Study of Nuclear Physics in China

【目的/意义】学者影响力评价是国内外研究的热点,本研究在学者影响力评价指标h指数与z指数中加入"主题"维度弥补过往研究的不足。【方法/过程】以核物理学科为例检验模型的有效性,利用网络爬虫在中国知网抓取文献共计11419条。经数据预处理后,首先,利用LDA模型对文献分类,得到我国核物理学科的八个研究主题:核反应研究、放射性衰变与新元素合成、核物理应用研究、核物理计算、核反应实验装置设计、核数据测量方法与数据库建设评价、核结构研究与核天体物理研究。其次,利用ATM模型对学者分类。最后,利用上述分类结果分别计算学者的hk指数与zk指数,并将结果与h指数和z指数比较。【方法/过程】经过相关性检验,研究结果与h指数和z指数有较高一致性。因此,利用LDA模型和ATM模型融合学者影响力评价指数h指数和z指数可以更加科学地评估学者的影响力,弥补了全学科评价指标的缺陷。

【Purpose/significance】Scholars’ influence evaluation is a hot topic at home and abroad. In this study, the "subject" dimension was added to the h-index and z-index of scholars’ influence evaluation indexes to make up for the deficiency of previous studies.【Method/process】The validity of the model was tested by taking nuclear physics as an example, and a total of 11,419 documents were captured on CNKI using network crawler.First of all, the eight research topics of nuclear physics discipline in China: the nuclear research, radioactive decay and new element synthesis, applied research in nuclear physics, nuclear physics experiment device design calculation, the nuclear reaction, nuclear data measurement and evaluation of database construction, the structure of the nuclear research and nuclear astrophysics research were obtained by LDA model of literature classification. Secondly, ATM model was used to classify scholars. Finally, the above classification results were used to calculate the hk index and zk index respectively, and the results were compared with h-index and z-index.【Result/conclusion】The correlation test results are consistent with h and z indexes. As a result, The integration of LDA model and ATM model with scholars’ influence evaluation indexes, h-index and z-index, can be used to evaluate scholars’ scientific research ability more scientifically, and make up for the defects of the whole subject evaluation index.