基于高阶概率权重矩的广义极值分布参数估计

Parameter Estimation of Generalized Extreme Value Distribution Based on Higher Probability Weighted Moments

研究基于高阶概率权重矩的广义极值分布参数估计。根据高阶概率权重矩法原理,建立了广义极值分布高阶概率权重矩估算参数模型。以陕北地区4个水文测站的年最大洪峰流量序列为例,结果表明:高阶概率权重矩法能赋予大洪水值更多的权重。蒙特卡洛试验表明:适当提高阶数可以减小误差,但阶数过高反而会增大误差。

This paper made research on parameter estimation of generalized extreme value distribution based on higher probability weighted moments. The model of higher probability weighted moments for estimating parameters of generalized extreme value distribution was established by the principle of higher probability weighted moments. The results of examples of annual maximum flow series in four stations indicate that the higher PWMs gives more weight to large flood values. The Monte Carlo experiments indicate that it is useful to raise the order appropriately.