Abstract: The online technology marketplaces is of great significance for the construction of an interconnected national technology trading network. From the perspective of knowledge transfer barriers caused by knowledge characteristics, it studied the industries characteristics of technology source of online technology marketplaces supported by Chinese government and social capital in this paper. It focused on the industry level factors related to technology marketplace and knowledge transfer, and studied how the above factors affect the possibility of industrial inventions entering different types of online marketplace. It attempted to prove that different types of online marketplaces have different preferences for industry selection, so as to make up the research gap of the attraction of online technology marketplace types to technology traders. 
   Based on the analysis framework of Dushnitsky and Klueter, it selected representative online platform enterprises according to certain principles, classified and counted their listed patents, and analyzed the influence of industry characteristic factors on the number of listed patents. Based on the types of patent transactions and the characteristics of online platform enterprises, it put forward the theoretical hypotheses of the barriers of knowledge transfer and the choice of online technology marketplace types. Two hypotheses were as follows: in the online technology marketplace of private enterprises, the industry knowledge fuzziness was obvious, which show that the industry knowledge fuzziness was positively related to the number of listed patents in the industry; in the online technology marketplace of state-owned enterprises, the industry knowledge was obvious, which show that the industry knowledge specificity was positively related to the number of listed patents in the industry.
   In the process of screening a typical online technology marketplaces, the first step was to use Baidu search engine, to conduct a combined search according to three groups of key search terms, to obtain a preliminary online marketplaces list and news reports of relevant marketplaces, and then preliminarily screen out some marketplaces that have not attracted media attention. The second step, according to the online marketplace activity and the requirements of the bilateral marketplace, was to exclude the platforms that have less listing and only serve the single type of institutions such as universities and governments. The third step, according to the classification requirements of the online marketplace for patents, it excluded the marketplaces that do not provide technology classification search, simple and broad classification standards, and/or search functions cannot be used. The fourth step, according to the operation status of the online marketplace, it excluded the stopped operation and newly established platforms. Finally, the determined online technology marketplace included two state-owned enterprise platforms and two private enterprise platforms. The steps to determine the industry classification were as follows: first, based on the national economic industry classification standard issued by the National Bureau of statistics of China, 96 categories of the industry were selected. The second step was to classify all the listed patents into the national economy industry category based on industry categories of the above four online technology marketplaces. In this study, the number of patents listed in the online marketplaces of different industries was the dependent variable; in terms of the selection of independent variables, the industrial linkage degree was taken as the proxy variable of industry knowledge fuzziness, and the industrial invention patent intensity was taken as the proxy variable of industry knowledge specificity. The negative binomial regression model was used for estimation, and the analysis software was stata14.1. 
    In the regression model of the number of patents listed in the online marketplace of private enterprise, although the regression coefficient of industry patent density is positive, the correlation is no longer significant, but the industrial linkage degree still maintains high significance, and the significance level of influence coefficient and sensitivity coefficient are 5% and 1%, respectively. In view of the operation flexibility and service diversification of private enterprise platform, it is more attractive for industries with high knowledge fuzziness to participate in its marketplace. In the regression model of the number of patents listed in the online marketplace of state-owned enterprises, the coefficient of industry patent intensity is positive and significant at the level of 5%; while the influence coefficient of industrial linkage degree becomes not significant, and the significance level of sensitivity coefficient also drops. It shows that in view of the marketplace stability, standardization and other characteristics of the online marketplace of state-owned enterprises, more industries with high requirements for knowledge specificity will be attracted to participate in its marketplace. The above results verify the two theoretical hypotheses. The results show that different types of online technology marketplace have different technology sources, and the barriers of knowledge transfer significantly affect the type preference of industry organizations to online technology marketplaces. The online marketplaces established with the support of social capital is better at reducing the knowledge fuzziness of the industry, while the online marketplace established with the support of the government is more attractive to industries with higher knowledge specificity. The theoretical significance of this study is as follows: firstly, from the perspective of knowledge transfer, it is proved that the technology sources of China′s online technology trading platforms have the characteristics of industry selection; secondly, under the national conditions of China, the online technology marketplace dominated by government and social capital has been formed, and the research proves that there are different industry selection preferences in the two different types of online marketplaces. This study appropriately responds to the call of theoretical research, and can increase the understanding of current technology trading platform enterprises in China. In China′s online technology marketplaces, especially in the initial stage of online technology transaction display and online contact between potential buyers and sellers, different types of platform enterprises play important and different roles. Significant differences in operation mode, personnel structure and service content of platform enterprises will lead to differences in platform technology sources, which also reflects the preferences of different technology owners for platform enterprises. 

Key words: industrial linkage, patent intensity, patent operation, intellectual property transaction, technology transfer