Why is it correct to use the sqrt to cancel out the sqrt operations in standard deviation?

2016-11-24 1:01 am
For a single number, sqrt(2^2)=2, we can get back the original number.

But for more than one number:
sqrt(2^2 + 3^2) = ~3.60 but not sqrt(2^2) + sqrt(3^2) = 2 + 3 =5

So I feel strange for standard deviation to use sqrt to cancel out the square operations.
更新1:

Should be "Why is it correct to use the sqrt to cancel out the *square* operations in standard deviation?"

回答 (1)

2016-11-24 3:21 am
Well, (a+b)^2 does NOT equal a^2 + b^2, rather it equals a^2 + 2ab + b^2.
Closely related, √(a+b) does NOT equal √a + √b, i.e. √ does NOT distribute over addition. Similarly, √(a^2 + b^2) does NOT equal a + b.

Example: √(3^2 + 4^2) means √(9+16) and since 9+16=25, you get
√(9+16) = √25 = 5. So, you would be wrong if you said √(3^2 + 4^2) = 3+4 because 3+4 = 7, and 7 does NOT equal 5.

This is just basic algebra and really has nothing to do with standard deviation.


收錄日期: 2021-04-18 15:50:17
原文連結 [永久失效]:
https://hk.answers.yahoo.com/question/index?qid=20161123170129AAOm538

檢視 Wayback Machine 備份