Abstract: 【Purpose/significance】As the main implementation carrier of scientific and technological innovation and talent training in my country, the research team studies the relationship between the characteristics of its cooperation network and innovation perfor⁃ mance, which will help to better realize the allocation and optimization of scientific research team resources and improve innovation performance.【Method/process】Taking the 1418 scientific research teams identified in all co-authored papers in the field of web of sci⁃ ence medical informatics from 2011 to 2020 as the research objects, from the perspective of the overall scientific research team, the scientific research teams are clustered based on the filtered features and the characteristics of each type of scientific research team are analyzed. The team's innovation performance affects the path.【Result/conclusion】The results of the study found: Based on the charac⁃ teristics of the cooperation network, the scientific research team can be divided into four types: two-person cooperation, streamline, network and hybrid, and the characteristics of the cooperation network of each type of scientific research team are significantly differ⁃ ent. In addition to the two-person cooperative team, in other types of scientific research teams, the smaller the network density, the easier it is to produce higher innovation performance【. Innovation/limitation】Based on data evidence combined with data mining meth⁃ ods, we effectively capture the nonlinear and complex relationship between the configuration effect of cooperative network characteris⁃ tics and innovation performance. The decision-making path of innovation performance is presented for providing practical reference for the improvement of research team construction and improvement of research performance.