Skip to content

moving distance between two sets of data

6 messages · David Winsemius, White, William Patrick

#
On the surface this seems pretty simple, but I flummoxed. I have two sets of numbers they bounce around zero, positive one and negative one. They have a relationship between them, where one diverges away from the other. I want create a second set of numbers that tracks that divergence. 

#Lets make some data like mine, kinda
Firstset <- runif(100, min = -1 , max =1)
Secondset <- runif(100, min = -1 , max =1)

#So something like:
Divergence <- abs (Firstset - Secondset)

#but this doesn't work because when Firstset is at .5 and Secondset is at  -.25 it returns .25 instead of .75

#another possibility is:

Divergence <- abs (Firstset) - abs (Secondset) 

#but when Firstset is at .5 and Secondset is at -.5 it returns 0 instead of 1

#It seems like there is a better way to do this. Any ideas?
#
On Aug 19, 2012, at 12:04 PM, White, William Patrick wrote:

            
abs( .5 - (-.25) ) should NOT return .25 so you need to produce a  
better example or point to specifics in the example you offered. If  
what you wanting what you are getting, then use set.seed(123) and  
refer to specific values.

 > abs( .5 - (-.25) )
[1] 0.75
#
On Aug 19, 2012, at 12:58 PM, David Winsemius wrote:

            
I meant to write:  "If you are not getting what you are wanting .... "
David Winsemius, MD
Alameda, CA, USA
#
The first method described produces a value other than the desired output every time the Firstset value is positive and the Secondset is negative, such as the second instance when seed is set to 123. The second method described produces a value other than the desired output every time the Firstset is negative and Secondset is positive, such as the first value derived from seed 123.
#
Also it occurred to me that my initial explanation was not explicitly clear as to what the desired output is. What I am trying to get is a moving absolute deviation between the two sets of numbers. This is not to be confused with the mean absolute deviation, or the median absolute deviation which are both something different and not what i am after.
#
On Aug 19, 2012, at 4:34 PM, White, William Patrick wrote:

            
The phrase "a moving absolute deviation" admits of several  
interpretations. I suggest you post the correct answer for some simple  
cases or that you be more mathematical in your description (as is  
suggested in the Posting Guide.)

set.seed(123)
  X <- sample(-5:5, 10)
  Y <- sample(-5:5, 10);

 > X
  [1] -2  2  5  4  1 -5 -3  3 -4  0
 > Y
  [1]  5 -1  1  4 -5  0 -4  3 -2  2
 > abs( tail(X,9) - head(Y,9) )
[1] 3 6 3 3 0 3 7 7 2

So this is c( abs(X[2] -Y[1]), abs( X[3]-Y[2], .....)
Again. Not a clear description (of what you do do not want), given  
that the problem involves two vectors.
David Winsemius, MD
Alameda, CA, USA