[R] which.minimums not which.min

[Philippe Grosjean]

What Fred is looking for is local minima/maxima, also known as turning 
points, or pits/peaks in a series.  You can look at ?turnpoints in 
pastecs library.

 > x <- c(1:4,0:5, 4, 11)
 > x
  [1]  1  2  3  4  0  1  2  3  4  5  4 11
 > tp <- turnpoints(x)
 > summary(tp)
Turning points for: x

nbr observations  : 12
nbr ex-aequos     : 0
nbr turning points: 4 (first point is a peak)
E(p) = 6.666667 Var(p) = 1.811111 (theoretical)

   point type       proba      info
1     4 peak 0.100000000 3.3219281
2     5  pit 0.002380952 8.7142455
3    10 peak 0.005952381 7.3923174
4    11  pit 0.666666667 0.5849625
 > plot(tp) # Only useful for a longer and more complex series!
 > # Get the position of peaks
 > (1:length(x))[extract(tp, no.tp = FALSE, peak = TRUE, pit = FALSE)]
[1]  4 10
Warning message:
arguments after the first two are ignored in: UseMethod("extract", e, n, 
...)
 > (1:length(x))[extract(tp, no.tp = FALSE, peak = FALSE, pit = TRUE)]
[1]  5 11
Warning message:
arguments after the first two are ignored in: UseMethod("extract", e, n, 
...)
 > # By the way, there are warnings although it works well (I ask on R-Help)

Now, you can easily code your which.minima() function using turnpoints:

x <- c(1:4,0:5, 4, 11)
x
tp <- turnpoints(x)
summary(tp)
plot(tp) # Only useful for a longer and more complex series!
# Get the position of peaks
(1:length(x))[extract(tp, no.tp = FALSE, peak = TRUE, pit = FALSE)]
(1:length(x))[extract(tp, no.tp = FALSE, peak = FALSE, pit = TRUE)]
# By the way, there are warnings although it works well (I ask on R-Help)

which.minima <- function(x) {
	if (!require(pastecs)) stop("pastecs library is required!")
	x <- as.vector(x)
	(1:length(x))[extract(turnpoints(x), no.tp = FALSE, peak = FALSE, pit = 
TRUE)]
}

which.minima(x)

Of course, you could optimize this code. This is just a rough solution!
Best,

Philippe Grosjean

..............................................<°}))><........
  ) ) ) ) )
( ( ( ( (    Prof. Philippe Grosjean
  ) ) ) ) )
( ( ( ( (    Numerical Ecology of Aquatic Systems
  ) ) ) ) )   Mons-Hainaut University, Pentagone (3D08)
( ( ( ( (    Academie Universitaire Wallonie-Bruxelles
  ) ) ) ) )   8, av du Champ de Mars, 7000 Mons, Belgium
( ( ( ( (
  ) ) ) ) )   phone: + 32.65.37.34.97, fax: + 32.65.37.30.54
( ( ( ( (    email: Philippe.Grosjean at umh.ac.be
  ) ) ) ) )
( ( ( ( (    web:   http://www.umh.ac.be/~econum
  ) ) ) ) )          http://www.sciviews.org
( ( ( ( (
..............................................................

Marc Schwartz (via MN) wrote:
> On Wed, 2006-03-15 at 11:32 -0800, Fred J. wrote:
> 
>>  Hi
>>   
>>  Is there a function which determines the location, i.e., index of
>>the all minimums or maximums of a numeric vector.
>>  Which.min(x) only finds the (first) of such.
>>   
>>  > x <- c(1:4,0:5, 4, 11)
>>  > x
>>   [1]  1  2  3  4  0  1  2  3  4  5 4 11
>>  > which.min(x)
>>  [1] 5
>>  > which.max(x)
>>  [1] 11
>>  >
>>   
>>  but I need 
>>  which.min(x)  to be 5 11
>>  which.max(x) to be 4 10
>>   
>>  thanks
>>   
> 
> 
> There is something wrong with your example code versus data here, since:
> 
> 
>>x
> 
>  [1]  1  2  3  4  0  1  2  3  4  5  4 11
> 
> 
>>which.min(x)
> 
> [1] 5
> 
> 
>>which.max(x)
> 
> [1] 12
> 
> 
> There is one one minimum value of 0 in that vector and only one maximum
> value of 11.
> 
> If you had a vector 'x':
> 
> 
>>x <- c(1:4, 0:5, 4, 0, 5)
> 
> 
>>x
> 
>  [1] 1 2 3 4 0 1 2 3 4 5 4 0 5
> 
> 
> You could then do the following to get the indices of the multiple
> min/max values:
> 
> 
>>which(x == min(x))
> 
> [1]  5 12
> 
> 
>>which(x == max(x))
> 
> [1] 10 13
> 
> 
> The only other thing that I can think you might be considering would be
> local minima/maxima in the vector and if that is what you want using:
> 
>   RSiteSearch("local minima")
> 
> or
> 
>   RSiteSearch("peaks")
> 
> 
> should lead you to some solutions that have been discussed previously.
> 
> HTH,
> 
> Marc Schwartz
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
> 
>