next up previous contents
Next: PEPSYS general photometry Up: Examples Previous: Period analysis

Comparison of two stochastic processes

Let the two tables OBSERVA.tbl and OBSERVB.tbl contain two sets of observations. Each set is stored in the DOUBLE PRECISION columns :TIME, :VALUE and :VAR containing the times of observation, data value and their variances.

 
 CREATE/GRAPHICS                   		 ! Create graphics window

SET/CONTEXT TSA ! Enable TSA package

NORMALIZE/TSA OBSERVA :VALUE V ! Normalize variance in both light

NORMALIZE/TSA OBSERVB :VALUE V ! curves to the same value of 1

COVAR/TSA OBSERVA OBSERVA AUTOCOVA 1. 0.1 24 LOG

Compute autocov. of `A'

PLOT/TAB AUTOCOVA :LAG :COVAR ! Plot autocov. function of `A'

COVAR/TSA OBSERVB OBSERVB AUTOCOVB ? ? ? LOG

Compute autocov. of `B'

PLOT/TAB AUTOCOVB :LAG :COVAR ! Plot autocov. function of `B'

COVAR/TSA OBSERVA OBSERVB CROSSCOV ? ? ? LOG

Compute crosscov. of `A' and `B'

PLOT/TAB CROSSCOV :LAG :COVAR ! Plot crosscovariance function

! Now you have to fit a common analytic formula to both autocor-

! relation functions, AUTOCOVA and AUTOCOVBB. The MIDAS FIT package

! or any other suitable tool may be used for this purpose.

! Choose one of the predefined function forms or code your own

! function UR, 0 < i < 10, in FORTRAN. Then, the analysis

! of the delay can proceed:

DELAY/TSA OBSERVA OBSERVB CHI2LAG 0 5 200 EXP 0,1,-0.25

! Do Chi2-time lag analysis

PLOT/TAB CHI2LAG :LAG :CHI2 ! Plot the results



next up previous contents
Next: PEPSYS general photometry Up: Examples Previous: Period analysis



Rein Warmels
Mon Jan 22 15:08:15 MET 1996