Abstract:    Universities, as the supplier of patent technology in the process of patent technology transfer, are responsible for patent knowledge creation and utilization. Enterprises are on the demand side and also the producers of products and services. Chinese government has successively issued a number of policies to coordinate the factors of various parties, promote the construction of bridges between suppliers and demanders, and ensure all participating entities in a sound and reasonable operation during the transfer process. The commercialization of scientific and technological achievements is one of the five construction tasks in the construction of "double first-class" university project. Moreover, technology transfer performance has also been an important index in the evaluation of world-class universities in China. Therefore, the establishment of robust evaluation indicators and the effective evaluation of patent technology transfer is essential to objectively reflect the performance and output of patent technology transfer work in universities, and it is also the basis for the prosper development of university technology transfer.
   In the early research of performance indicator system, researchers mainly focused on the evaluation of the benefits of technology transfer, such as the number of patents licensed or assigned and the actual revenue brought by them. Subsequent research started to concentrate on the potential of technology transfer, such as the number of papers published, the number of patents disclosed, amount of research funding acquired, and the number of personnel equipped, etc. Gradually, evaluation began to cover the whole process of technology transfer as much as possible: from scientific discovery, invention disclosure to patent evaluation; from the creation and management of intellectual property to operation; from individual behavior, group consciousness to organizational management. The cut-in from different angles has enriched the existing evaluation research, but also made the empirical operation more and more difficult, the indicators are difficult to aggregate and lack of systematicness. There are some common limitations in the existing research: First, the sample size of universities is limited. Second, the selection of variables is similar, especially domestic research output indicators are in a simple structure, consisting of the number of technology transfer contracts and the actual income from technology transfer. Third, the research perspective is relatively single and the research ideas are relatively fixed. 
    In view of the above limitations, this paper uses patent data as the new data source to build multi-dimensional performance indicators to enrich and supplement the existing research. It has the following advantages: first, it has a wide data dimension, the transfer of patent rights can be surveyed in the patent transaction information, and the basic patent information contains a variety of information, such as the transferor, inventor, the description of the technology, the recipient, etc., that makes it possible to depict a complete transfer path of patent technology. Secondly, it can be traced back and all statistical indicators can be traced back to a certain patent transfer transaction.
    The research methods of this paper are as follows: (1) Construct a patent basic information table, and extract the patent transfer information from the unstructured transaction data to construct the university patent technology transfer table which concludes the information such as application number, original patentee, transferor, receiver, date of transfer, transferor address, recipient address, transfer type, etc. (2) Join the two tables by primary key patent application number, the entire data includes the inventor, inventor information, patent information, transfer path, recipient and their address which are necessary to support the research in this paper. (3) Based on the above data, construct the performance evaluation index of university patent technology transfer from the four basic elements of knowledge transfer. (4) Comprehensively use evaluation indicators to evaluate technology transfer performance from various levels. The entropy method is used to calculate the corresponding weight of each index, and the Jenks best natural breakpoint method is used to classify the comprehensive value. (5) According to the first level indicators, the top 30 universities are analyzed by hierarchical cluster analysis to indicate the common characteristics of universities in different categories.
   In this study, 17500 pieces of directly-related data are screened out. Based on the theory of knowledge transfer, the performance index system is constructed from four aspects: the transferor, the receiver, the characteristics of technical knowledge and the path of technology transfer, which makes up for the shortcomings of the existing indicators. At the same time, the performance of patent technology transfer of 238 universities in China is evaluated by using the index system. Shanghai Jiaotong University, Tsinghua University, Southeast University are in the first echelon; Harbin University of Technology, Jiangnan University and other universities are in the second echelon; Chongqing University, Wuhan University and other universities are in the third echelon. Nevertheless, the performance forms of different universities can be different within the same echelon. In the empirical process, there are obvious differences in the number of independent transfers, cooperative transfers and internal transfer as well as the length of transfer path among universities. This is related to the internal technology transfer mode of universities, and fully reflects the differences in the management of technology transfer in universities in China. Universities in the north emulate with each other to establish remote institutions in the southeast part of China, aiming to cooperate with local industries in personnel training, innovation and entrepreneurship, achievement commercialization, etc. Meanwhile, universities in the north lack third-party intermediary institutions, besides the intellectual property service industry is not active. Thus, the northern universities mainly rely on the internal asset management department in school. On the contrast, the direct transfer mode in the southeast is dominant, the intellectual property service industry is very active, and the number of third-party service institutions is obviously large. As a whole, among the four levels, there is little difference in terms of technical knowledge characteristics, which is not optimistic, revealing the lack of hierarchical management of intellectual property. Therefore, treating different patents differently and accordingly optimizing the allocation of resources are the next steps for deepening patent technology transfer.

Key words: university patent technology transfer, patent data, indicator system, performance