基于FA-DEA聚类的模式分析方法及在我国软件园研究中的应用

A Development Pattern Analysis Method Based on FA-DEA Clustering with its Application in China Software Parks Research

模式分析方法日益成为现代经济管理研究的主流方法之一,但是目前关于模式分析的结构化方法的研究还不多见.本文提出了基于FA-DEA聚类的模式研究结构化方法,这个方法区别于以往文献的是采用了混合聚类方法,综合运用关键因素的绝对量信息和效率的相对量信息.基于上述方法,本文确定了我国软件园发展的三种发展模式.并对三个模式的条件、特征以及它们之间的分界与演进做了详细的描述.最后,给出了若干建议.

As development pattern analysis is becoming more and more popular in modern economics and management research, a structural method to do development model research is presented in this paper. It consists of three key steps: model classification; model comparison; and model evolution analysis. Especially, two points of the method are different from other researches, that is, using DEA results to represent Intuitional Efficiency and Scale Efficiency; using a ″two-steps method″ to utilize absolute metrics on key factors and relative metrics on efficiency comprehensively in the clustering model. This paper also gives a detailed application example of this method in China Software Parks Development Model Analysis. In this example, we got the results fairly good and convincing, which approves the effects of the method we put forward in this paper.