基于关联规则挖掘的研发团队识别方法研究

The Research of Core R&D Team Recognition Method Based on Association Rules Mining

摘要:

针对当前单个人才引进不能很好地满足企业、行业和区域日益升级的人才需求问题，探寻一种新的方法来从专利数据中识别某一机构的核心研发团队，从而为企业、行业和区域批量引进人才提供决策支持。由于传统共现分析方法存在缺陷，研究引入并优化关联规则挖掘中的FP-Growth算法来实现核心研发团队的识别，并以中国科学院长春光学精密机械与物理研究所为对象验证方法的有效性，最后以中国科学院为研究对象进行实证研究。实证研究挖掘得到中科院7支核心研发团队，包括中科院海洋研究所的海洋生物制品课题组等。

Abstract:

This paper explored a new method of recognizing the core R&D team of an institution in order to provide support for talents introduction of enterprises, industries and regions, as single talent introduction cannot meet the increasing demand of talents. Since traditional co-occurrence method has many drawbacks, we introduced and optimized FP-Growth algorithm of association rules mining to recognized core R&D team from patent data. Then, Changchun Institute of Optics, Fine Mechanics and Physics of CAS was selected as experimental subject to test the method's validity. Finally, an empirical study of CAS was conducted. Result of empirical study indicated that CAS has 7 core R&D teams, including Marine Biological Products Research Group from Institute of Oceanology of CAS and so on.