Abstract: In order to achieve a high-precision law-based business environment assessment and improve the performance of the support vector machine (SVM), the improved grey wolf optimization algorithm(IGWO)is applied to the legalized business environment assessment SVM model, this paper establishes the IGWO-SVM model of business environment evaluation based on 13 two-level evaluation indexes of business environment of rule by law, Business Law Enforcement Environment, business judicial environment and business compliance environment. Classification accuracy(ACC) , Sensitivity(Sensitivity), Specificity(Specificity)and Carl Friedrich Philipp von Martius correlation coefficient(MCC)were compared among GA-SVM, PSO-SVM and GWO-SVM, the IGWO-SVM model has higher precision, efficiency and stability, which provides a new method for business environment assessment.

Key words: rule of law business environment, support vector machine, grey wolf optimization algorithm, index system, reverse learning

摘要： 为实现高精度的法治化营商环境评价,提高支持向量机(Support Vector Machine,SVM)模型的评价性能,将改进的灰狼算法(Improved Grey Wolf Optimization Algorithm,IGWO)应用于优化SVM模型的法治化营商环境评价,建立基于营商法治环境、营商执法环境、营商司法环境和营商守法环境4个方面构建出的13个二级法治化营商环境评价指标的法治化营商环境评价IGWO-SVM模型。通过比较GA-SVM、PSO-SVM和GWO-SVM的分类准确率(ACC)、灵敏度(Sensitivity)、特异性(Specificity)以及马修斯相关系数(MCC)可知,法治化营商环境评价IGWO-SVM模型具有更高的精度、效率和稳定性,为法治化营商环境评价提供了新的方法。

关键词: 法治化营商环境, 支持向量机, 灰狼优化算法, 指标体系, 反向学习