首页 > 解决方案 > 检查两个矩阵的对应元素之间是否为零

问题描述

我有两个矩阵代表下置信区间和上置信区间。我想检查所有元素的两个对应元素之间是否为零。需要 TRUE/FALSE 输出。

下限:

P_diff_lw <- structure(c(NaN, 0.0781621066919929, 0.0113707247330811, 0.00484959866281143, 
0, NA, 0, NaN, -0.469268662772997, 0.211358332924958, -0.601772116208517, 
-0.462536625853965, NA, -0.0939534951944804, NaN, -0.183798369715781, 
-1.18450010949587, 0.171270462047824, 0.0772117721799035, NA, 
-0.0957898935363418, NaN, 0.0135099896535079, -0.959164384110204, 
-0.964777156674447, -0.566058830346901, NA, 0.00250882014261652, 
NaN, 0.0316883050829714, -0.478120520974224, 0.0881622360514131, 
-0.367959997198539, NA, -0.696986043305979, NaN, -0.00196587987960171, 
-0.000500855512583966, 0.0545913669650907, -0.14886050041401, 
NA, 0.242738490542487, NaN, 0, -0.00329569932572391, -0.00106553727189175, 
0.0382307407644255, NA, -0.461668123968531), .Dim = c(7L, 7L), .Dimnames = list(
    i = c("Aaa", "Aa", "A", "Baa", "Ba", "B", "Caa"), j = c("Aaa", 
    "Aa", "A", "Baa", "Ba", "B", "Caa")))

上限:

P_diff_up <- structure(c(NaN, 0.0944267257953167, 0.070596488381673, 0.0252255893071134, 
0, NA, 0, NaN, -0.300854614964494, 0.239461339206189, 0.417561589892728, 
0.260782239889053, NA, 0.0939534951944804, NaN, 0.334632307351749, 
1.07442985188463, 0.430233297350672, 0.291209280451675, NA, 0.0957898935363418, 
NaN, 0.270753969737355, 0.666424337271797, -0.0540198358067556, 
0.425707953153919, NA, 0.197491179857383, NaN, 0.0698345375566225, 
0.282570169686167, 0.272740019587685, 0.148661751584504, NA, 
0.0969860433059792, NaN, 0.0628795854633581, 0.0496811833814364, 
0.125859760854458, 0.157632430238572, NA, 0.490594842790846, 
NaN, 0, 0.0196891419486747, 0.0311407252418166, 0.145979785551364, 
NA, 0.128334790635197), .Dim = c(7L, 7L), .Dimnames = list(i = c("Aaa", 
"Aa", "A", "Baa", "Ba", "B", "Caa"), j = c("Aaa", "Aa", "A", 
"Baa", "Ba", "B", "Caa")))

标签: rmatrixconfidence-interval

解决方案


正如评论指出的那样,P_diff_lw <= 0 & P_diff_up >= 0这里的工作(除了 NAs)

在更一般的情况下,如果您有两个数字a并且b 不知道哪个更大,您可以仅计算当且仅当其中一个数字为正而另一个为负时a*b<=0哪个将返回 true


推荐阅读