Keywords - Function groups - @ A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Library: times
See also:

Quantlet: jump
Description: detects jump points in the time series. Optional parameter alpha controls the the sensitivity of the procedure. Recomended values are 0.5 ... 4.0. The default value is 2.0. The output vector j, which has the same length as data, indicates the detected jumps. NaN values are not allowed.

Usage: j = jump(data {,alpha})
Input:
data n x 1 matrix
alpha 1x1 matrix (scalar)
Output:
j n x 1 matrix

Example:
library("times")   ; load library
length = 200       ; size of the data
t=1:length
greed = t/length   ; greed on the interval [0,1]
vr = 0.3          ; variance of the white noise
;create a signal on the unit interval with
;two jump points at the 0.4 and 0.7
signal = 0.5.*greed.*(greed <= 0.4)+(0.4+0.2.*greed).*(greed>0.4)
signal = signal + 0.4.*(greed>0.7)
randomize(123456789012)
data = signal+ vr.*normal(length) ;add white noise to the signal
alpha = 4;
jm = jump(data,alpha)
; plot the data
; the observable data will be shown by the blue crosses
; red line corresponds to the signal
; magenta line shows the locations of the detected jumps
signal=t~signal
jm= t~jm
data = t~data
setmaskl(signal,t',4,1,0)
setmaskp(signal,0,0,1)
setmaskl(jm,t',5,1,2)
setmaskp(jm,0,0,1)
setmaskl(data,t',0,2,1)
setmaskp(data,1,11,8)
output = createdisplay(2,1)
show(output,1,1,data)
show(output,2,1,jm,signal)
setgopt(output,2,1,"title","signal and detected jumps")
setgopt(output,1,1,"title","noisy signal")

Result:
the display "output"



Author: Y. Golubev, 19980609
(C) MD*TECH Method and Data Technologies, 05.02.2006