基于引用时序特征的唤醒专利识别研究

A research on identification of awakening patents based on citation time series features

唤醒专利是特殊类型的专利,由于技术内涵的超前性,在生命周期的早期阶段,处于罕有引用的休眠状态;在经历多年的沉寂后,随着技术发展和市场变化,最终受到关注并获得高引用而被唤醒;作为大器晚成的专利类型,唤醒专利尚未受到足够重视,亟待开展深入的研究工作。选取无线通信网络领域的专利作为研究样本,基于专利引用的时序特征分析,研究构建唤醒专利的识别方法;运用该识别方法筛选唤醒专利,剖析其典型特征,研究提出利用唤醒专利特征,分析和监测相关技术的发展趋势。

The awakening patent is a special type of patents.Due to its advanced technical connotation,the awakening patent was in a dormant state that received rare citations in the early stage of its life cycle.After years of silence,with the technology development and market changes,it woke up due to attention and high citations.Therefore,we can also call it the sleeping beauty patent.As a"late bloomer"type of patent,the awakening patent has not received enough attention.It is urgent to carry out in-depth research.In order to identify awakening patents that hidden in a large number of patents effectively,we analyze the data of patent citation time series to reveal the time series features of patent citation.We select patents in the field of wireless communication network as research samples.The analysis of patent citation time series shows that the citations of most patents show an inverted U-shaped distribution.That is,the number of citations increases first and then decays.However,there are significant differences between the features of citation time series of few patents and the overall patents.These very few patents had only sporadic or even zero citations in their early stage.However,they finally attracted attention and gained high citations after years of silence,showing the distinct feature of"late bloomer"and reflecting the essential feature of awakening patents.We propose the identification method of awakening patent based on the"late bloomer"time series feature of patent citation.Firstly,we construct the awakening patent index API based on the analysis of the time series features of patent citation.That is,the ratio of the patent citations after the 11th year to the total citations.Then,we rank the patent samples in descending order according to the total citation to filter patents ranked in the top 20%of the total citation as highly cited patents.Finally,we filter patents with API≥0.8 one by one among the highly cited patents.They are the awakening patents in the patent sample.Then we utilize the awakening patent index API constructed by the research to filter the awakening patents in the target technology field.We filter out 36 awakening patents in total.The statistical results show that awakening patents only account for 2.35%of the total number of 1534 samples and 11.7%of the 307 highly cited patents.This shows that the awakening patent is a rare patent type,which may be the important reason that awakening patents remained unheeded for a long time.Then,we select the typical awakening patent to analyze the sleeping cause and awakening process of awakening patents.The research shows that awakening patents are in a dormant state of rare concern in the early stage of the life cycle.With the development of technology and the change of demand,the involved technology begins to receive attention.It reflects in the rapid growth of patent citation and the awakening of long-dormant patents.Therefore,awakening patents are concerned in the latter stage of life cycle,which actually reflects the technology development trend and the transfer of technology hotspots.We can utilize the feature of awakening patents and the above identification method to filter awakening patents timely.Thus,to monitor the development trend of relevant technology by identifying and analyzing awakening patents in the target technology field.