专利情报应用问题及改进方法

Application Problems and Improving Methods of Patent Intelligence

[目的/意义]针对因专利系统结构和行为特征而出现的不良性状问题,可利用专利价值指数进行改进。[方法/过程]首先,讨论引文、授权、专利族、续费和对抗五种基础性指标的应用条件和参数设定;其次,根据欧洲专利局的数据集对以上指标进行检验,结果显示百分位数排序方法能够纠正相异分布参数问题;最后,在泊松分布的假设下利用简单算术平均数复合可计算专利价值指数。[结果/结论]对比结果显示,敏感于技术发展的专利价值指数自20世纪90年代以来快速下降,而专利申请倾向却保持上升,因此与专利申请强相关的情报并不能合理解释技术发展水平。

[ Purpose/Sigulficance ] For solving the bed characters problems resulted from the structure and behavior features of patent system, patent value index can been an ideal solution. [ Method/Process] Initially, the paper discussed the applying condition and parameter setting of five basic indicators, including patent citation, granted decisions, patent family, renewal and opposition. Then, according to the test resuk on the indicators with the European Patent Office dataset, it was concluded that percentile ranking function can solve the hetero- geneity problem of the distribution parameters. Finally, on the hypothesis of Poisson distribution, the patent value index was calculated by simple arithmetic weight. [ Result/Conclusion] The comparison result suggested that for patent value index, which is sensitive to the de- velopment of the technology, there has been a sharp fall since the 1990s, however the patent application propensity has kept rising. There- fore, it is not appropriate for the intelligence strongly connected to patent application to be used in explaining the technology development level in a reasonable way.