摘要
对于小样本的时间序列应用2种改进的PWM方法,分别用修正的Bootstrap方法和随机加权法对样本进行再抽样,然后根据PWM方法进行区间估计,并以潘家口水库径流量为例,对2种改进的PWM方法与传统方法进行比较,改进的PWM方法的精度更高,在相同的置信水平下,置信区间更短.
For using the improved PWM, we exponentially down - weight the abnormal points to alleviate their influence and dynamically adjust the weights to the different data to gain more robustness. Meanwhile,we get a shorter confidence interval with the improved PWM than other estimation methods. So in the condition of small sample, we can use the improved PWM method of interval estimation. The two improved PWM methodsare the random weighting method and the improved bootstrap method respectively. We take PanjiakouReservoir as an example to compare the effects of the two improved PWM methodswith those of the traditional methods. The result shows that the improved PWM methods have high precision, and the confidence interval is shorter at the same confidence level.
出处
《云南民族大学学报（自然科学版）》
CAS
2015年第4期300-303,共4页
Journal of Yunnan Minzu University:Natural Sciences Edition
基金
江苏省水利科技创新基金(2011059)
河海大学自然科学基金(2009426311)
