[R] red noise

bongi guido.bongi at iro.pg.cnr.it
Mon Jun 6 19:02:04 CEST 2005


help


Dear sirs,I'd like to add a red noise spectra to a periodogram:


#given z as time series of 61 values
>z
[1] 10500.0000 12044.6667 13589.3333 12222.0000  9132.6667  7000.0000
 [7]  7000.0000  5464.6667  2735.1733  1048.3200 18843.0667 38235.2533
[13] 37907.4800 19892.2533  4269.0533  4524.4000  5889.5133  5756.3000
[19]  2924.4000  2527.2000  2011.0000  1054.2000   127.4000  7005.6000
[25] 14607.6000 27384.0000 45817.1000 54225.6000 39942.5000 25961.3667
[31] 22606.0000 12321.1333  7731.2667 23948.4000 39990.2000 25766.3333
[37]   473.6000   788.8933  1131.9000  3209.7800  6968.5000  8114.1667
[43]  4855.4000 10016.6667 10944.9333 11314.0000 12890.2667 13107.4667
[49]  6314.0000  2320.6667  2277.0000  4142.0000  7451.6667  8884.9333
[55]  7292.0000  1407.0000  1142.5333  7927.2000  8964.1333  2380.9227
[61]    28.0000



> s3 <- spec.ar(z) # estimate(plot) spectral density
> s3$spec # retrieves 500 points of powerdensity
> s3$freq #and their f
> plot(s3$freq,s3$spec,type="l") # produces another plot of s3, but i was
unable to find a method for red noise

Guido Bongi CNR ISAFM PERUGIA ITALY




More information about the R-help mailing list