摘要
专利收益具有动态性、随机性、不平衡性、相对性和隐蔽性等特征,是专利创造、运用、保护、管理和服务等的核心,与专利数量、质量和价值等关系密切。以中国授权、国内居民拥有的法定保护期于2019年届满的专利为研究对象,基于逻辑回归-BP神经网络模型对专利收益影响因素分析发现:专利主体特征对专利收益的影响程度最大,专利客体特征和专利事件特征对专利收益的影响程度相对较弱,且比较接近;主体特征中专利权人类型、专利权人所在地区经济状况和发明人数对专利收益的影响程度依次递减;客体特征中同族专利数、独立权利要求数和专利引证数对专利收益的影响依次递减;事件特征中专利审查时间、诉讼涉案数和专利被引数对专利收益的影响程度依次递减。
The patent profit is the benefits obtained by the patentee or related beneficiaries based on the exclusive right of the patented technology, and its legal basis is the valid patent. The patent profit is the core of patent creation, application, protection, management and service, which is closely related to patent quantity, quality and value. The patent profit has the characteristics of dynamic, random, unbalanced, relative and hidden. The subject’s scale and type, nationality and geographical location, patent operation ability and litigation ability, patent preference and strategic level affect the patent profit;the object factors such as the length and intensity of patent, the family size of patent, the technical status of patent, the validity and certainty of patent affect the patent profit;patent litigation, patent invalidation, patent examination and other events also affect the patent profit. This paper puts forward three preliminary methods for quantifying the patent profit, which are based on thepatent maintenance time, application price, and official awards. The maintenance time quantifying method is used in this research, which regards maintaining patent as a strategic behavior of a rational economic man. Only when the expected profit brought by the patent issatisfactory, the patentee will choose to maintain the patent. Therefore, the patent maintenance time can be regarded as a measure of the patent profit. Taking patents granted by China and owned by domestic residents whose legal protection period expires in 2019 as the research object, based on the logistic regression-BP neural network model, the empirical analysis results of the influencing factors of patent profitareas follows. First, based on the perspective of patent subject, compared with patents owned by private patentee, public patentee has relatively lower patent profit, while mixed patent owners have relatively higher patent profit. Compared with patents in economically developed regions, the profit of patents in general and poor economic regions is significantly lower. Second, based on the perspective of patent object, compared with the patents with single independent-claim, the probability that the patents with multiple independent-claim are profitable patents is 1.248 times. Compared with low family-number patents, the probability that high family-number patents are profitable patents is 2.651 times. Third, based on the perspective of patent event, compared with non-litigation patents, the probability of litigation-patents being profitable patents is 3.087 times. The probability of patents with longer examination time becoming profitable patents is 1.376 times that of patents with shorter examination time. In general, the patent subject characteristics have the greatest influence on the patent profit;the patent object characteristics and the patent event characteristics have a relatively weak influence on the patent profit and are relatively close;the influence degree of the patentee type, the economic status of the patentee region, and the inventor number in the patent subject characteristics on the patent profit orderly decreases;the influence degree of the patent family number, the independent claim number, and the patent citation in the patent object characteristics on the patent profit orderly decreases;the influence degree of the patent examination time, the litigation case number, and the patent cited number in the patent event characteristics on the patent profit orderly decreases.
作者
乔永忠
刘睿
QIAO Yong-zhong;LIU Rui(Intellectual Property Research Institute of Xiamen University,Xiamen 361005,China)
出处
《科学学研究》
CSSCI
CSCD
北大核心
2021年第10期1850-1859,共10页
Studies in Science of Science
基金
国家自然科学基金面上项目(71874148)。
关键词
专利收益
影响因素
逻辑回归
BP神经网络
patent profit
determinants
logistic regression
BP neural network