CAJ | 학술논문

提出了一种面向最大值指标的广义拉依达准则粗差处理方法，为最大值指标下粗大误差的有效鉴别提供了参考依据。该方法假设观测序列服从独立同分布的正态分布，从最大观测值被误作为粗差数据的风险分析入手，对拉依达准则的判定标准进行了改进，推导并给出了广义拉依达准则的粗差判决条件。实践应用的结果表明，该方法是可行的。
제출료일충면향최대치지표적엄의랍의체준칙조차처리방법，위최대치지표하조대오차적유효감별제공료삼고의거。해방법가설관측서렬복종독립동분포적정태분포，종최대관측치피오작위조차수거적풍험분석입수，대랍의체준칙적판정표준진행료개진，추도병급출료엄의랍의체준칙적조차판결조건。실천응용적결과표명，해방법시가행적。
A maximum-error specification oriented gross error identification method based on general Paǔta criterion is proposed, which provides a reference for gross error identification in maximum-error specification. It is assumed that the target stochastic observa-tion sequence is subject to IID normal distribution. Then, through a risk analysis on mistaking the maximum observation value as the gross error data, some modifications are made to the classic Paǔta criterion, and the general Paǔta criterion is introduced. The gross error identification threshold calculation method is also given. Practical application test results show that the method is feasible.