割合(2項確立)に対する信頼限界の近似

概要

論文の詳細を見る
For obtaining approximate binomial confidence limits, the most commonly used is p^^^±u_α√<p^^^(1-p^^^)/n>, where p^^^=x/n, n is the sample size, x is the number of defectives, and u_α denotes the upper α point of the standard normal distribution. This approximation is simple and its meaning is intuitively clear, but its accuracy is too poor to meet practical requirements unless n is very large. To improve its accuracy, it is proposed to use p^^^*=(x+d)/(n+e) instead of p^^^=x/n, where two constants d and e are suitably chosen. Numerical investigations are made to compare the accuracy of the new method with those of other procedures.
社団法人日本品質管理学会の論文
1980-10-15