Abstract:    The analysis of the evolution law of patent knowledge absorption and diffusion in the field of medicine in China and the influence of knowledge absorption on knowledge diffusion will help us select and absorb relevant resources quickly, provide data support for increasing the intensity of investment in basic research, and provide support for the strategy of integrating scientific and technological resources. This paper takes medicine, one of the six research fields classified by the National Bureau of Economic Research, as the research object, and 2326 patents and it′s cited 43822 patents and 35573 papers in the United States Patent and Trademark Office in the field of Chinese medicine from 2000 to 2014 as the data source. First, the trend of the absorption and diffusion of patent knowledge with time is analyzed; second, the 2000-2014 time period is divided into three five-year periods, the changes of mean value and standard deviation of each measurement index of knowledge absorption and diffusion in the three five-year periods is analyzed, and the multivariate analysis method of variance is used to determine whether each index is significantly different over time; finally, the knowledge absorption measure is used as the independent variable, the knowledge diffusion measure is used as the dependent variable, and zero-inflated negative binomial regression and multiple linear regression are used for statistical analysis. It provides data support for increasing the intensity of basic research investment, provides basis for the integration strategy of science and technology resources in China, and provides reference for the formulation of effective achievement transfer mechanism.
     The results show that (1) the absorption of patent knowledge increases year by year, it shows that more and more patented technologies refer to basic research knowledge and existing patented technologies in their research and development. But the diffusion of knowledge does not form a stable trend, it shows that although the number of patents in the field of medicine in China is increasing year by year, the quality of patents still needs to be strengthened; (2) the number, novelty, and knowledge components of patent literature absorbed by patents increase over time, but the quality of patent literature absorbed by patents declines over time. The numbers, knowledge components of paper literature absorbed by patents increase over time, and the novelty generally increases, but the quality of papers absorbed by patents also declines over time. The patent diffusion rate becomes faster, but the scope of diffusion becomes narrower and the diffusion intensity gradually weakens. Using Pillai′s trajectory, Wilks′ Lambda, Hotelling trajectory, and Roy′s maximum as test statistics, multivariate analysis of variance is used to compare whether the various measures of patent knowledge absorption and diffusion change significantly over time, the result show that the measurement indicators of knowledge absorption and diffusion are significantly different in different time periods; (3) the quantity and quality of patent literature knowledge absorption and the quality of paper literature knowledge absorption have a negative impact on patent knowledge diffusion. This is because unlike paper knowledge absorption, patent knowledge absorption is a product of competition between competitors for market share, while patents have novelty, so the more patents are absorbed, the greater the threat to the novelty of the patent. At the same time, learning basic research results will not immediately benefit, and will consume time and energy, so that its energy investment in technology development will be reduced, so it is not conducive to the validity of patents; (4) the absorption of paper knowledge components has a positive impact on patent knowledge diffusion. This is because if the number of knowledge components of papers in a patent absorption is more, which means that these papers cover a wider range of disciplines, then the patents that absorb these papers have a stronger knowledge base and innovation compared with other patents. Such a cycle may spread to other broader patents more quickly with a greater chance; (5) the number of absorbed papers is negatively correlated with the breadth and intensity of knowledge diffusion. This is because the more the number of absorption papers, the more time the inventors need to understand and absorb the knowledge, so the smaller the breadth and intensity of diffusion; (6) the absorbed patents knowledge components have a positive impact on the breadth of knowledge diffusion, but not with intensity and speed of knowledge diffusion, for reasons similar to the fourth point analysis above; (7) the novelty of knowledge absorption in papers and patents have no effect on the breadth, intensity and speed of knowledge diffusion. This is because knowledge diffusion does not require the timeliness of knowledge absorption. There are many classic papers published for a long time and patents with earlier publication date may be established technical achievements. 
    On the contrary, in recent years, the papers and patents have stood on the shoulders of giants and rely on more systematic and mature theoretical foundation and method technology, there may be more outstanding achievements, so the novelty of knowledge absorption has no significant impact on knowledge diffusion. Therefore, (1) the absorption of patent knowledge limits the scope of the right of new patents to a certain extent and weakens the novelty of new patents, so the quality of the patent itself cannot be judged solely by the quantity and quality of cited patents in the future; (2) it is more important to mine the tacit knowledge involved in the absorption of paper literature, and the absorption of explicit and tacit knowledge requires a lot of time and energy, so the impact of the quantity and quality of cited papers on the quality of patents needs to be observed and analyzed from a wider time window; (3) knowledge components represent the scope of patent technology and the scope of the disciplines of paper, in the context of interdisciplinary background, the more knowledge components are absorbed, the more new methods and technologies can be brought, which may break through the existing thinking patterns and technical limitations, and increase the possibility of technological innovation. Therefore, in the process of patent knowledge absorption, attention should be paid to the absorption of multi-technical scope and multi-disciplinary knowledge; (4) the diffusion of patent knowledge is not related to the timeliness of knowledge absorption, so it is not necessary to consider too much its novelty by referring to previous knowledge and technological achievements, as long as it can be used as the knowledge base and technology reserve of new invention, it can be used as a part of knowledge absorption.

Key words: patent, paper, knowledge absorption, knowledge diffusion, multivariate analysis of variance, zero-inflated negative binomial regression, multiple linear regression