基于专利文本挖掘的产业关键共性技术识别与应用研究

Research on Industry Key General Technologies Identification and Application Based on Patent Text Mining

摘要:

破解“卡脖子”技术难题，需要科学识别产业关键共性技术、准确判断产业关键共性技术发展现状。通过构建LDA主题模型、计算关键性得分和共性得分、运用相似度分析确定阈值等步骤，来识别产业关键共性技术。随后运用该方法，以高校和上市公司专利数据为样本，来评估我国产业关键共性技术水平。研究发现，无论是高校还是上市公司，属于关键共性技术专利的比例均不高，申请专利中二者的比例分别为14.7%、10.2%，授权专利中二者的比例分别为13.8%、10.1%，高校专利占比略高于上市公司。进一步分析高校和上市公司的典型关键共性技术，发现得分位居前列的技术主题与当前广受关注的关键共性技术非常一致，表明该方法是有效、可取的。科学判断产业关键共性技术水平，能够为政府制定针对性的创新支持政策提供决策参考。

Abstract:

In order to deal with stranglehold problems in key core technologies, it is necessary to identify industry key general technologies scientifically and judge the development status of industry key general technologies accurately. The article identifies industry key general technologies by constructing LDA topic models, calculating criticality scores and commonality scores, and applying similarity analysis to determine thresholds and so on. The methods were then applied to assess the level of industry key general technologies in China, using patent data from universities and listed companies as samples. It was found that the proportion of patents belonging to key general technologies was not high for both universities and listed companies, with the proportion of patents applied for by universities and listed companies belonging to key general technologies being 14.7% and 10.2% respectively, and the proportion of granted patents belonging to key general technologies being 13.8% and 10.1% respectively, with universities accounting for a slightly higher proportion than listed companies. Further analysis of the typical key general technologies of universities and listed companies showed that the top scoring technology topics were very consistent with the current key general technologies of wide interest, indicating that the method is effective and desirable. Judging the level of industry key general technologies scientifically can provide a reference for the government to make decisions on targeted innovation support policies.