【作者】 万小丽；

【导师】 朱雪忠； 田力普；

【作者基本信息】 华中科技大学 ， 技术经济及管理， 2009， 博士

【摘要】 全球范围内专利数量的急剧膨胀引起各国对专利质量的关注,人们意识到专利质量在不断下降。专利质量过低不仅降低竞争力,遏制创新,引起众多的权利纠纷,还增加社会成本,浪费社会资源。很多发达国家的专利战略已经从关注数量调整为提高质量,我国才刚刚进入重视专利质量的新阶段。重视专利质量离不开评价专利质量,而评价专利质量的核心又是专利质量指标问题。本文首先将专利质量界定为专利技术对使用者形成竞争力的重要程度,为专利质量指标研究奠定基础。目前国内外对专利质量尚无统一定义。现有的专利质量定义要么只判断专利是否符合专利授权实质性条件,要么仅仅考虑法律效力、技术质量和经济效益中的某一个因素。本文认为,专利的本质是技术,专利技术的利用最终要转化成产品,能否使专利使用者具有竞争优势是检验专利技术质量的市场标准。因此,本文从竞争力的角度对专利质量进行界定,使其兼具法律性、技术性和经济性,成为一个广义的、相对的概念。其次,本文从理论上设置了专利质量指标,并对其进行效力验证。现有的专利质量指标纷繁复杂,本文从专利引证、专利维持、专利范围和其他四个角度系统地分析了现有指标的含义、原理和效力,并对有缺陷的指标进行修正或重新设计,遴选出较为合理的指标,包括专利参考文献数量、非专利参考文献数量、自引科学论文数量、被引次数、科学关联度、技术循环周期、扩散指数、“授权后第8(5)年是否维持”、专利族大小或“是否存在专利族”、权利要求数量、技术覆盖范围和发明人数量。随后,本文选用中国获奖专利和中国在美国获取的专利作为样本,首先运用均值差异法对个别指标的效力进行验证,除去与专利质量关系不大的指标;然后运用逻辑回归分析方法从整体上验证剩余指标的效力,最终确认被引次数、权利要求数量和发明人数量三个有效指标。最后,本文构建了专利质量指数模型,并用其进行专利质量评价实例分析。基于三个有效指标,本文首先用逻辑回归的标准回归系数确定指标权重,然后用固定效果法将指标原始数据转换成相对指数,最后用线性加权综合法构建专利质量指数模型。该指数可以评价单项专利的质量和多项专利的整体质量,而且不受专利的授权年份和技术领域的约束;不仅可以横向或纵向地比较评价对象的专利质量差异,还可以判断评价对象的专利质量水平在系统中所处的相对位置。此外,本文还利用该模型进行了例证分析,结果显示该模型具有可适用性。更多还原

【Abstract】 While the quantity of patents is rising sharply in the world, many countries realize a trend that the patents’ quality is falling. Low-quality parents could lose competitive strength, hold back innovation, encourage more charges of infringement and litigation, also could increase social costs and waste social resources. Many developed countries have transformed their patent strategies from quantity to quality. But our country just recently began to attach important to patent quality. This shift to patent quality has set a question that how to evaluate patent quality and which indicators should be used.This paper first defines patent quality as the importance of patent’s technology to form users’ competitive strength, which provides fundamental base for the further research of patent quality indicators. At present, there is no uniform definition of patent quality at home and aboard. Some definitions of patent quality consider how well the patent meets the statutory substantive requirements particularly while others only emphasize one factor of legal validity, technology quality and economic perspective. This paper holds that the nature of patents is technology itself which should finally be transferred into patent products, and so the market standard of measuring patents’ technology quality is how well the patent will make the user have a competitive advantage. Therefore, this paper defines patent quality from view of competitive strength. This definition is a broad and relative concept which focuses on all factors of legal, technology and economy.Secondly, this paper theoretically sets up patent quality indicators and verifies their efficacy. As the existing indicators is complicated, this paper divides them into four fields about patent citation, patent renewal, patent scope and others, and analyzes their meaning, principle and efficacy. This paper also finds some defective indicators and attempts to amend them or design new ones. This paper chooses some reasonable indicators, including patent references, non-patent references, self-backward paper citations, forward citations, science linkage, technology cycle time, generality index, "whether renew at the 8^th (5^th ) year after patent grant", patent family size or "whether have a patent family", number of claims, technology scope and number of inventors. As for these indicators, based on samples of Chinese award-winning patents and patents that China got from USPTO, this paper first verifies the efficacy of some individual indicators through mean comparison and removes the indicators that have no relation to patent quality, and then verifies all of the rest indicators through logistic regression and finally confirms three efficiency indicators, including forward citations, number of claims and number of inventors.Finally, this paper structures a model of patent quality index and gives examples for measuring patent quality with it. Based on the three efficiency indicators, this paper regards standard regression coefficients in logistic regression as weight of indicators, transfers initial data of indicators into relative index by the way of fixed-effect, and structures the model of patent quality index with linear weighted comprehensive assessment method. The index can evaluate quality of single patent and multiple patents, and free from grant year and technology field. Furthermore, the index not only can horizontally or vertically compare the differences of patent quality of evaluated objects, but also can judge the relative level of patent quality of evaluated objects in the system. In addition, this paper gives examples for measuring quality of single patent and multiple patents with the index, and demonstrates the applicability of the index.更多还原