Research Of Patent Information Of Domestic Pathological Diagnosis AI Industry

[Purpose] Policy preference and technological progress have promoted the rapid development of domestic pathological diagnosis AI industry in recent years,but the state has not issued relevant overall planning.At present,the disordered and complex development of this field has increased a lot of information costs for researchers.To solve this problem,this study attempts to take patent mining as the basis,aiming at(1)showing the development context and technology layout status of the industry through visual tools;(2)Discuss the technical competition and cooperation situation of the whole industry and the main patentees through the quantitative analysis model;(3)Provide information science support for relevant researchers and investors,and practice the application of information analysis theory in specific scenarios.[Methods] The data of this research came from soopat Chinese patent database.After data cleaning,the following analysis was carried out:(1)the co-occurrence network analysis method was used to study the citation information such as patentees andkeywords,sort out the cooperation relationship,related technology,patent timeline,invalid status source and highly cited patents;(2)the emergence detection analysis method was used to study the IPC classification number,and calculate the emergence intensity value and emergence time of subdivision technology,to analyze resource inclination and hot spot technology.(3)The patent quality index method was used to calculate the patent quality ranking of major patentees,and analyze the similarities and differences between them and the patent number ranking.(4)the technology life cycle method was used to calculate the technology growth rate,technology maturity coefficient and new technology characteristic coefficient of colleges and universities and enterprises,so as to analyze their technical competitiveness,current competition situation and future development expectation.[Results] The annual growth rate of related patents was 27.4%,and the proportion of rejected and revoked patents was 41.6%.There was no overall cooperation network,and cooperation was limited to universities and enterprises.The network model was star type,ring type and bus type.There is no honeycomb or tree type.The network diverged around the high-yield core authors.The keyword co-occurrence network had no absolute core,and some keywords appeared more than 5 times but existed independently.High cited patents were concentrated in large enterprises and well-known universities.Tencent had the highest patent quality index of 0.712,and Lanting,Shanda,Zhejiang University and Ping An were 0.694,0.657,0.612 and 0.591 respectively.Various enterprises and universities had differentiated distinctive high-profile IPC classification numbers.The overall emergence intensity was > 20.75% of the high-profile IPC classification numbers were in G06,and 90% of the emergence time was later than 2012.The technology growth rate in this field was > 0.2,the technology maturity coefficient was > 0.5,and the new technology characteristic coefficient was > 0.8.The technology maturity coefficient of enterprises reached the peak earlier than that of universities.[Conclusion] This field is technology intensive and relevant research should be done well in patent early warning to avoid duplication.There is huge room for improvement in researcher cooperation,and more technical exchanges should be carried out.There are many small directions that can be studied in the field of pathological diagnosis AI,and there is no obvious boundary.Different research objects have high emergent strength IPC classification numbers associated with their own attributes and technology accumulation.In the fierce competition,they have differentiated their fields of expertise.The industry as a whole is in the period of technological growth.At the same time,the period of technological growth has just begun.The research of colleges and universities is generally earlier than that of enterprises,but their ability to convert research results into products is lower,the overall competition situation is more and more intense,and the competitiveness of various research objects is obviously differentiated.