【摘要】 在测量数据为正态分布时 ,多用莱特准则剔除异常数据 ,但应注意莱特准则应用的条件是测量次数n需满足n≥ 10 ;在测量数据为均匀分布时 ,在工程测量中 ,剔除异常数据可采用 1 73σ(X)作为判据 ,文章就此作了理论论证。应当指出 ,均匀分布是电测仪器仪表测量数据中的一种极为重要的分布 ,文章就此作了深刻的说明。显然 ,讨论工程测量中均匀分布的测量数据中异常数据的剔除 ,有着现实的工程意义更多还原

【Abstract】 This paper discusses the elimination of abnormal data in engineering measurement, points out that when measuring data distribute in different situation (distribute in abnormal distribute or distribute in uniform distribution), the rules of eliminating data are different (in first situation the eliminating rule is 3σ(X), in second situation the eliminating rule is 1.73σ(X)),and gives the theoretical demonstration. This paper also points out that uniform distribution is a very important distribution in enginerring measurement.更多还原