专利引文的幂律分布研究

Power Law Distribution of Patent Citation

[目的/意义]针对专利引文的统计分布研究,探讨专利引文的统计分布特征。[方法/过程]以美国专利商标局的申请专利为数据来源,在德温特专利数据库中下载2006-2015年间所有存在引用的美国专利,应用统计指标和回归分析分别探索这些专利在不同国际分类专利代码部中的统计特性与分布特征。[结果/结论]结果表明,专利被引频次具有类似于复杂网络无标度的分散性,并且专利被引频次的概率分布遵循幂律分布。

[Purpose/Significance]Focusing on the statistics of patents citation,the paper explores the statistics characteristics of patent citation.[Method/process]All the cited patents that applied to United States Patent and Trademark Office over the period 2006-2015 are downloaded from the Derwent Innovations Index.Then the statistical indicators and regressions analysis are used to reveal the distribution characteristics of the cited patents that are divided by the different categories of International Patent Classification Code.[Result/Conclusion]The results demonstrate that the decentralization of patent citation counts is similar to the scale-free in complex networks and the probability distribution of patent citation follows power law.