基于高效用关联规则挖掘的技术创新合作团队发现——以新能源汽车产业专利数据为例

以新能源汽车产业专利数据为例，将专利被引用频次、同族数量作为衡量专利原始效用的主要指标，并通过两步骤的高效用关联规则挖掘算法，获得专利权人的高效用合作项集，以发现高价值的技术创新合作团队。实证表明，将专利价值的评价模型与高效用关联规则挖掘方法相结合，能够极大提升技术创新合作团队发现的召回率。此外，与传统关联规则挖掘方法相比，基于高效用关联规则获得的高价值团队在分布数量、分布类型等方面呈现明显优势。

Technology Innovation Cooperative Team Discovery Based on Efficient Association Mining

Taking the patent data in the new energy vehicle industry as an example, the frequency of patent citation and the number of the same family were taken as the main indicators to measure the original utility of the patent. Using the two-steps effective utility association rule mining, both cooperative itemsets of patentee with effective utilities and high-value technique innovation cooperative teams were mined. Experiments indicated that the recall rate of discovering technique innovation cooperative teams was enhanced largely by combining the patent value evaluation model with the effective utility association rule mining. Furthermore, the high-value teams identified by the effective utility association rule mining showed obvious advantages when compared with the classical association rule mining in terms of distribution quantity and distribution type.