Abstract: In recent years, many patents have been sold through auctions in China. Patent aucitons absorb substantial interest from a diverse collection of companies and individuals and reflect a continuing trend toward development of a market for patents. However，the literatures on patent auction and its sale rate are very rare. By analyzing a dataset of 89 patents of first patent auction held by China Technology Exchange on December 16 of 2010, we identify different characteristics of auctioned and non-auctioned patents and the determinants of sale rates of patent auctions by using the Logistic regression model. The results show pricing way and such patent specific attributes as number of inventors, type of patent, number of main classes have a significant effect on the sale rates of patent auctions.

摘要： 近年来，许多专利技术在我国以拍卖的方式得以交易。专利拍卖吸引了众多企业和个人的关注和兴趣，并日益成为专利技术市场的一种发展趋势。但是，目前我国研究专利拍卖及其成交率的文献还很少。以中国技术交易所2010年12月16日举办的第一届专利拍卖的89件专利为样本，通过双样本t检验比较拍卖成交专利和未成交专利的特征差异，并利用Logistic回归模型分析专利属性和拍卖方式等因素对专利拍卖成交率的影响。研究结果表明，专利拍卖的定价方式和专利属性中的发明人个数、专利类型、专利分类数对专利拍卖成交率有显著影响。

关键词: 专利拍卖, 中国技术交易所, 成交率, Logistic回归模型, patent auctions, China Technology Exchange, sale rates, Logistic regression model