本文从区域创新产出能力、区域创新投入、区域创新环境水平等三个方面，构建了区域创新能力的评价指标体系，并提出了基于Fussy-GRNN的区域创新能力评价模型。进而，本文在分析分布密度（SPREAD）选取原则的基础上，利用该模型进行了区域创新能力评价的实证分析。本文的研究表明，“基于Fussy-GRNN的区域创新能力评价模型”具有较强的自学习能力、高度的容错性和鲁棒性等、小样本适应性、训练速度快、模糊处理等特点，证明了该模型的合理性和有效性。

This paper constructs the new index system for regional innovation capability evaluation from three aspects such as regional innovation output capability, regional innovation input capability, the level of regional innovation environment. Then this paper proposes the regional innovation ability evaluation model based on the Fussy-GRNN. Then, on the basis of the analysis of the selection principle for distribution density(SPREAD) of GRNN, the paper makes an empirical analysis of regional innovation capacity evaluation by using the model. This study shows that the model has advantages such as good self-learning ability and nonlinear mapping ability, high fault tolerance and robustness, adaptability of small sample, faster training speed, fussy processing etc. This study also proved that the model of regional innovation capacity evaluation based on Fussy-GRNN is rational and valid.

国家社科基金项目“复杂动态环境下产业集群创新中的群体知识协同行为与机制研究”（13CGL012）；山东理工大学青年支持计划项目“动态环境下的产学研协同创新机制研究：基于微观视角的实证研究”（2014）；山东省高校人文社科计划项目“基于知识协同的产业集群创新模型研究”（J13WF03）