Question

所以，我在R中有以下代码：

y
a <- -0.1    
test <- (1/((y+as.numeric(!y))*(a-1)))  
test  
test^a  
-0.9090909^a

给我输出：

> y  
 [1] 0.00000000 0.06024096 0.00000000 0.01098901 0.00000000 0.00000000
 [7] 0.01829268 0.00000000 0.06976744 0.00000000 0.04380665 0.01351351
[13] 0.00000000 0.00000000 0.00000000 0.00000000 0.00310559 0.00000000
[19] 0.00000000 0.00000000 0.09957447 0.00000000 0.03738318 0.00000000
> a <- -0.1
> test <- (1/((y+as.numeric(!y))*(a-1)))
> test
 [1]   -0.9090909  -15.0909091   -0.9090909  -82.7272727   -0.9090909
 [6]   -0.9090909  -49.6969697   -0.9090909  -13.0303030   -0.9090909
[11]  -20.7523511  -67.2727273   -0.9090909   -0.9090909   -0.9090909
[16]   -0.9090909 -292.7272727   -0.9090909   -0.9090909   -0.9090909
[21]   -9.1297591   -0.9090909  -24.3181818   -0.9090909  

> test^a  
 [1] NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN
[17] NaN NaN NaN NaN NaN NaN NaN NaN  

> -0.9090909^a  
[1] -1.009577

我不明白为什么在测试电源时它给了我NaN，显然，我可以简单地从测试中复制一个值并将其提升到电源a，这样就可以了。使用test [1] ^ a也不起作用。

Answer 1

-0.9090909 ^ a与-(0.9090909 ^ a)相同。请注意括号。

但是，您的test包含负值，而您无法取负数的根。请尝试(-0.9090909) ^ a进行验证：

> (-0.9090909) ^ -0.1
[1] NaN

Answer 2

你正试图取一个负数的根。如果您希望进行复杂的算术运算，则需要转换为复数

as.complex(test) ^ a
# [1] 0.9601644-0.3119763i 0.7249946-0.2355650i 0.9601644-0.3119763i
# [4] 0.6115633-0.1987090i 0.9601644-0.3119763i 0.9601644-0.3119763i
# [7] 0.6435367-0.2090978i 0.9601644-0.3119763i 0.7357171-0.2390490i
#[10] 0.9601644-0.3119763i 0.7022627-0.2281790i 0.6243418-0.2028609i
#[13] 0.9601644-0.3119763i 0.9601644-0.3119763i 0.9601644-0.3119763i
#[16] 0.9601644-0.3119763i 0.5389643-0.1751201i 0.9601644-0.3119763i
#[19] 0.9601644-0.3119763i 0.9601644-0.3119763i 0.7623605-0.2477059i
#[22] 0.9601644-0.3119763i 0.6912151-0.2245894i 0.9601644-0.3119763i

R的奇怪行为

2 个答案: