I have attached a text file representing the centralized amplitude of a signal,
sampled at 30Hz, whose length N = 6922
My goal is to remove the trend. I am using package "simsalabim".
I ran command decompSSA with L = length(Amps)/5
The reason is that I have SSA/MTM toolkit running in Mac/OS. SSA/MTM
documentation, relative to SSA, recommends that N/10 <= L <= N/5.
Documentation of simsalabim package recommens that L=N/2
I ran the command getSignal and felt uncomfortable at picking values for
parameters C0 and r0.
I would appreciate some suggestion/guidelines because these values seem to
influence the trend extraction.
I chose C0=0.0005, r0=0.0005 which are both below the Nyquist frequency
1/Amps.dc$L = 0.0007223346
As a consequence, getSignal finds two Eigenvectors capturing the trend.
>From my experience with toolkit SSA/MTM I know this output can be misleading.
I am trying to figure out whether trend has really been found by
double-checking the print-out data:
> Amps.signal$trend
k smoothness explainedVariance
1 1 0.134495414 26.254096
11 12 0.006693601 1.627475
The above print-out marks EOF_1 and EOF_12 as capturing all the trend in the
signal.
I canNOT understand the difference between "smoothness" and "explainedVariance"
I tried to compare such values with the leading frequency associated with EOF_1
and EOF_12.
Unluckily, I am unable to draw any comclusion:
> Amps.dc$freq[1] #EOF_1 LEADING FREQUENCY
[1] 0.0007225434
> Amps.dc$freq[12] #EOF_12 LEADING FREQUENCY
[1] 0.002890173
"getSignal" on-line documentation states about Trend "smoothness":
The proportion of variance explained by the frequencies <=q omega0.
But nowhere I could find what "q" is equal to. I guess the above sentence
indicates a fraction of the nyquist frequency Omega_0 ????
"getSignal" on-line documentation states about Trend "explainedVariance":
Percentage of the variance explained by the Eigenvalues with rank k
Does the above variance refer to the signal variance ????
I would greatly appreciate your help at using this package properly.
Thank you in advance.
Maura
e tutti i telefonini TIM!
Vai su
e tutti i telefonini TIM!
Vai su
e tutti i telefonini TIM!
Vai su
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
