分数阶广义积分-微分方程Riesz基的置信域估计

Confidence Region Estimation for Riesz Based on Fractional Order Generalized Integral Differential Equation

分数阶广义积分-微分方程Riesz基的置信域估计关系到方程是否存在稳定解,由分数阶广义积分-微分方程的初值解构成Riesz基,采用广义最小二乘估计(GLS)方法构成Riesz基的回归参数的置信域,广义积分-微分方程局部解存在性根据广义特征函数的分数阶非线性增长性约束条件进行验证。在重特征值的根子空间中通过Lyapunov泛函分析分数阶广义积分-微分方程Riesz基的置信域,通过计算最小二乘估计(OLS)估计的经验覆盖概率提高置信域估计的精度。

The problems of the confidence domain estimation equation related to stable fractional generalized integral differential equation based on Riesz, a Riesz based by the initial value of generalized integral and the fractional differential equations, estimated using generalized least squares (GLS) regression parameter confidence domain based Riesz method, according to the existence of fractional order nonlinear generalized the characteristic function of the growth of the constraint conditions to verify local solutions of the generalized integral differential equation. The generalized confidence region of Integro differential equations of fractional order Riesz based Lyapunov functional analysis in the eigenvalues of the root subspace, by calculating the least squares estimation (OLS) estimation of the empirical coverage probability improve confidence domain estimation accuracy.