基于药物专利信息资源的新兴技术探测研究

Research on Emerging Technology Detection Based on Drug Patent Information Resources

本文的核心目的是及早识别哪些药物专利可能蕴藏着新兴技术。专利的前向引用反映了专利的市场价值以及技术新兴的早期信号,因此本文选择高被引专利作为新兴技术的代理指标,引入前向视角和后向视角构建主题原创性、文本创新性和文本独特性指标,同时选取先前研究中被广泛用于新兴技术预测的指标,形成一套基于药物专利的指标体系,然后使用负二项回归模型预测专利授权后5年的前向引用次数,最终通过逻辑回归模型进一步证实了基于药物专利的新兴技术指标体系对专利的前向引用具有较好的探测效果。

The main purpose of this paper is to identify which drug patents may contain emerging technologies as soon as possible.The forward citation of patent reflects the market value of patent and the early signal of emerging technology.Therefore,this paper chooses highly cited patents as the proxy indicator of emerging technology,introducing forward and backward perspectives to construct the indicators of topical originality,textual innovation and textual uniqueness,and selects the indicators widely used in the prediction of emerging technology in previous studies.A series of indicators based on drug patent is formed,and then the negative binomial regression model is used to predict the number of forward citations in five years after patent authorization.Finally,the logistic regression model is utilized to further confirm that the patent-based indicators of emerging technology have a good prediction effect on the forward citations of patents.