ggplot2 :: stat_ecdf（）中的意外终结点行为

Question

我有一些增益图表的数据。 （我对所有目标结果都有模型分数的百分位数。）

CKEDITOR.replaceAll(function (textarea, config) {
    if (textarea.classList.contains("ignore_me")) {
        return false;
    };
    <...>
});

在先前版本的public class Combination { private static int count; public static void main(String[] args) { String[] inputs = new String[] {"12345", "1234", "123", "12", "1"}; for(String input : inputs){ count = 0; System.out.print("output for " + input + " is:"); combination(input); System.out.println("\nCount for input:" + input + " is " + count); } } private static void combination(String input) { combination("", input); } private static void combination(String prefix, String input) { System.out.print(prefix + " "); count++; int n = input.length(); for(int i = 0; i < n; i++){ combination(prefix + input.charAt(i), input.substring(i + 1)); } } } 中，我能够构建一个看起来像这样的增益图表：

但是在红色圈出的端点处没有奇怪的行为（所需行为为蓝色）

使用以下声明：

> dput(data)
structure(list(obs_set = structure(c(1L, 1L, 2L, 2L, 2L, 2L, 
1L, 1L, 1L, 2L, 1L, 1L, 2L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
1L, 1L, 2L, 1L, 1L, 2L, 1L, 2L, 1L, 1L, 2L, 2L, 1L, 1L, 1L, 1L, 
1L, 2L, 1L, 2L, 2L, 1L, 1L, 2L, 2L, 2L, 1L, 2L), .Label = c("Training/Validation", 
"Test"), class = "factor"), pctl = c(0.999903481425, 0.99980696285, 
0.999584970127501, 0.999498103410001, 0.999488451552501, 0.999353325547502, 
0.998706651095003, 0.998610132520004, 0.998291621222504, 0.998204754505004, 
0.997615991197506, 0.997577383767506, 0.996284034862509, 0.99615856071501, 
0.99603308656751, 0.995724227127511, 0.995376760257512, 0.995328500970012, 
0.995038945245012, 0.994894167382513, 0.994855559952513, 0.994787996950013, 
0.993957937205015, 0.993948285347515, 0.993600818477516, 0.99185383227002, 
0.991670446977521, 0.990087542347525, 0.989672512475026, 0.98802204484253, 
0.987607014970031, 0.985531865607536, 0.982472226780044, 0.979624928817551, 
0.979219550802552, 0.977183008870057, 0.97584140067756, 0.975407067090062, 
0.975020992790062, 0.974654222205063, 0.965166446282587, 0.965040972135087, 
0.96396961595259, 0.962850000482593, 0.961662822010096, 0.9598579246576, 
0.95200131265262, 0.933730346405166, 0.836372059802909, 0.725665254278186
)), .Names = c("obs_set", "pctl"), row.names = c(NA, -50L), class = "data.frame")

有什么变化吗？我在ggplot2版本2.1.0，R版本3.2.4。我几个月前发誓这很好，但我最近升级了。