Statistical distributions
Statistical distributions
复制标题
DOI：
10.1145/586740.586743
复制
发表时间：
1983-07
期刊：
影响因子：
--
通讯作者：
R. E. Wheeler
中科院分区：
文献类型：
--
作者：
R. E. Wheeler
我们提出了常见的连续概率分布函数的概率积分和逆。概率积分是用级数或连分数来计算的，在Abramowitz和Stegun [I]中，大部分是用它们的方程号来确定的。足够的条款是用来确保完全准确的机器与18个十进制数字;列入额外的条款并不困难，因为明显的性质，系数的系列和连分数。通过牛顿迭代得到逆，当概率积分接近期望概率时终止迭代。
We present probability integrals and inverses for common continuous statisticaldistribution functions. The probability integrals are evaluated by series or continued fractions, identified for the most part by their equation numbers in Abramowitz and Stegun [I]. Sufficient terms are used to insure full accuracy on a machine with 18 decimal digits; the inclusion of additional terms is not difficult, because of the obvious nature of the coefficients in the series and continued fractions. The inverses are obtained by Newtonian iteration, which is terminated when the probability integral is nearly equal to the desired probability.