Find Peaks

Find peaks (maxima) in a time series.

findpeaks(x, nups = 1, ndowns = nups, zero = "0", peakpat = NULL,
          minpeakheight = -Inf, minpeakdistance = 1,
          threshold = 0, npeaks = 0, sortstr = FALSE)

Arguments

x

numerical vector taken as a time series (no NAs allowed)

nups

minimum number of increasing steps before a peak is reached

ndowns

minimum number of decreasing steps after the peak

zero

can be `+', `-', or `0'; how to interprete succeeding steps of the same value: increasing, decreasing, or special

peakpat

define a peak as a regular pattern, such as the default pattern [+]{1,}[-]{1,}; if a pattern is provided, parameters nups and ndowns are not taken into account

minpeakheight

the minimum (absolute) height a peak has to have to be recognized as such

minpeakdistance

the minimum distance (in indices) peaks have to have to be counted

threshold

the minimum

npeaks

the number of peaks to return

sortstr

logical; should the peaks be returned sorted in decreasing oreder of their maximum value

Details

This function is quite general as it relies on regular patterns to determine where a peak is located, from beginning to end.

Value

Returns a matrix where each row represents one peak found. The first column gives the height, the second the position/index where the maximum is reached, the third and forth the indices of where the peak begins and ends — in the sense of where the pattern starts and ends.

Note

On Matlab Central there are several realizations for finding peaks, for example “peakfinder”, “peakseek”, or “peakdetect”. And “findpeaks” is also the name of a function in the Matlab `signal' toolbox.

The parameter names are taken from the “findpeaks” function in `signal', but the implementation utilizing regular expressions is unique and fast.

See also

hampel

Examples

x <- seq(0, 1, len = 1024)
pos <- c(0.1, 0.13, 0.15, 0.23, 0.25, 0.40, 0.44, 0.65, 0.76, 0.78, 0.81)
hgt <- c(4, 5, 3, 4, 5, 4.2, 2.1, 4.3, 3.1, 5.1, 4.2)
wdt <- c(0.005, 0.005, 0.006, 0.01, 0.01, 0.03, 0.01, 0.01, 0.005, 0.008, 0.005)

pSignal <- numeric(length(x))
for (i in seq(along=pos)) {
  pSignal <- pSignal + hgt[i]/(1 + abs((x - pos[i])/wdt[i]))^4
}
findpeaks(pSignal, npeaks=3, threshold=4, sortstr=TRUE)
#>          [,1] [,2] [,3] [,4]
#> [1,] 4.972146  134  118  144
#> [2,] 4.960051  799  786  817
#> [3,] 4.590829  257  246  306

if (FALSE) { # \dontrun{
plot(pSignal, type="l", col="navy")
grid()
x <- findpeaks(pSignal, npeaks=3, threshold=4, sortstr=TRUE)
points(x[, 2], x[, 1], pch=20, col="maroon")} # }