摘要
经典的极大似然估计和抗差的广义极大似然估计都是将随机样本看成独立积事件，在积概率最大的准则下，求参数估值，前者不能抗拒粗差，后者能抗拒粗差。和极大似然估计是将随机样本看成独立和事件，在和概率最大的前提下，求参数估值，从理论分析和计算结果来看，和极大似然估计优于积极大似然估计。
In classical maximum likelihood estimate and robust general maximum likelihood estimate,random sample is thought of as an independent product event.In the presence of estimating parameters on the rule of the happening maximally probablity of product event,theformer can not restrict gross error,but the latter can do it.The sum maximum likelihood estimate,given by the author in this paper, considers random sample as an independent sumevent.The parameters are estimated on the condition of the happening maximally probablityof sum event.According to theory analysis and computing result,sum maximum likelihood estimate is better than product maximum likelihood estimate.
关键词
极大似然估计
估计准则
抗差
maximum likelihood estimate
product maximum likelihood estimate
summaximum likelihood estimate
