基于DEA和决策树的区域创新效率评价与分析

Abstract

科学、客观地评价与分析区域创新系统的效率是分析和制定创新政策的基础，数据包络分析(DEA)是评价区域创新系统(称为决策单元)相对有效性的常用评价方法。决策树模型是呈树形结构，表示基于特征对实例进行分类的过程，是一种应用广泛的可视化数据挖掘的重要工具。本文将结合DEA模型和决策树模型的各自优点，提出综合运用DEA与决策树方法的评价分析模式，在评价我国31个区域创新系统绩效的基础上，建立分类决策树并分析其在制定区域创新系统的战略决策(或政策制定)中的作用。 To scientifically and objectively evaluate and analyze the efficiencies of regional innovation systems is a good base to analyze and make policies of regional innovation. Data Envelopment Analysis (DEA) is one of the most used methods for evaluating the efficiency of regional innovation systems (which are referred to Decision Making Units). Decision tree model is a tree structure, which is a process of classification based on feature; it is an important, useful and visual tool of data mining. Combining the advantages of DEA and decision tree model, this paper presents a hybrid model to evaluate and analyze the efficiencies of regional innovation systems using DEA and decision tree method. Based on the evaluation of the efficiencies of 31 regional innovation systems in China, the classification decision tree is established and its role in strategic decision making (or policy making) of regional innovation systems is analyzed.