基于整体性思想的数据分析

Data Analysis Based on Global Thinking

通过对风险函数的分解,首先得到了传统上在独立同分布模型或线性模型假定下被忽视的度量复杂系统非线性程度的重要指标—扭曲度,然后对如何降低风险进行了讨论.逐步引出了干扰度、偏离度和信息分解比,它们与风险函数和扭曲度一起组成五个指标;最后权衡这五个指标,达到对稳定中心度量指标的控制.这也是用整体性思想进行数据分析的一种尝试.

Through the decomposition of risk function, we first get an important indicator for the measure of nonlinearity of complex systems which has be traditionally neglected in the independently and identically distributed models or linear models. Then how to reduce risk is discussed. The we gradually introduce the departure degree, the disturbing degree and the information-decomposition ratio which constitute the five indicators together with the distortion degree and the risk function. Finally, we discuss how to control measure indicators of the stability center. It is also used to do a try for data analysis based on global thinking.