We calculated the matching values of technology supply and demand texts based on texts semantic similarity with Word2Vec and Cosine similarity algorithms, and then proposed a new index named Supply-Demand Matching Efficiency (SDME) to measure the matching efficiency of online technology trading platforms (OTTPs). Through the empirical research on the three types of OTTPs, the findings are as follows: First, the SDME of Zhejiang Market (Government-Owned, Government-Operated, GOGO), Technology E Market (Government-Owned, Contractor-Operated, GOCO), and Keyi Market (Market-Owned, Market-Operated, MOMO) are 64.69%, 54.38% and 28.99% respectively, indicating that the government plays an important role in attracting effective technology suppliers and demanders to participate in online trade and standardizing information expression, thereby improving the SDME. Second, by comparing the SDME and the newly announced signing rate of each OTTP, we found that the OTTP with high SDME also has high signing rate, and the changing trend of the two is consistent. Third, we used the TextRank and Latent Dirichlet Allocation (LDA) to study the topic distribution of technology supply and demand, and calculated the topic differences of each OTTP, which are 70%, 75%, 84% respectively. The Technology E Market and Zhejiang Market have low topic differences and high SDME, while Keyi Market has high topic differences and low SDME, which indicated that the topic differences have a negative effect on SDME. Intuitively, measuring the semantic matching efficiency of supply and demand texts on OTTPs can help the suppliers and demanders to retrieve information accurately, and assist the OTTPs to carry out trade promotion and evaluate trade performance.

我们利用Word2Vec和Cosine相似度算法基于文本语义相似度计算技术供需文本的匹配值，然后提出一个新的索引，称为供需匹配效率（SDME），以衡量在线技术交易平台（OTTPs）的匹配效率）。通过对三种类型OTTP的实证研究，得出以下结论：首先，浙江市场（政府所有，政府运营，GOGO）的SDME，技术E市场（政府所有，承包商运营，GOCO）和Keyi Market（市场所有，Market-Operated，MOMO）的比例分别为64.69％，54.38％和28.99％，这表明政府在吸引有效的技术供应商和需求者参与在线贸易和标准化信息表达方面发挥着重要作用，从而改善SDME。其次，通过比较SDME和每个OTTP的最新发布签名率，我们发现SDME高的OTTP也具有很高的签名率，两者的变化趋势是一致的。第三，我们使用TextRank和潜在Dirichlet分配（LDA）来研究技术供求的主题分布，并计算出每个OTTP的主题差异，分别为70％，75％，84％。技术电子市场和浙江市场的主题差异低，SDME高，而可可市场的主题差异高，SDME低，这表明主题差异对SDME有负面影响。直观地，测量OTTP上供求文本的语义匹配效率可以帮助供需者准确地检索信息，