Stochastic resonance - page 34
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
So what is the red line in Fig.2? The difference of energies of the real and imaginary parts? Very interesting. And what do the numbers in the value area of the indicator mean?
I can not answer yet, unfortunately, I am waiting for a rebuttal to my question about the library. The 'FFT Fast Fourier Transformation Library' because I could not understand what is outputted there
make it a separate variable altogether
P.S. Perhaps I mention the 'Spectral Analysis' indicator too often :), but the amplitudes work properly there - you can just take from it.
Simply removing the 0 frequency in this way is a hatchet job, although it does show where you need to go next. There are lateral lobes, they remain. We should definitely apply a Hemming window, as an option (but it's bad with energy, like a tractor) we should look at different windows when we get there.
I think to do variant of the following, as we know frequency interfering us (0 market does not move) and response function of each filter sin(x)/x. We should carefully calculate and subtract all sidewalls from all filters.
After removing the sidewalls, reverse Fourier transform (or convolution), remove trend function of the y=a+bx type and apply Hemming window, while applying direct Fourier transform again.
Now plot everything on a graph Signal and noise energy, before removal 0, after removal, after trend removal + output a and b coefficients. Then I think we will have a tool that allows us to investigate the market.
How do you like it?
Looks like there is an error in the library, or my hands are crooked again :( posted my question here 'FFT Fast Fourier Transform Functions Library') if anyone can check it. Am I right or not. Try it in matlab to check it.
In short, you are feeding different data to the fft input of the matcad and to the fastfouriertransform input from klot. Please don't be offended, but I've advised you 2 times to follow the link from the library's header http://alglib.sources.ru/fft/ and figure out the input and output data format for the functions, but you clearly haven't done that. This is the third and last one. By the way, the format is different for each function.
Simply removing the 0 frequency in this way is a hatchet job, although it does show where you need to go next. There are side lobes, they remain.
If you sum the amplitudes by starting to look at frequencies from hmax, you cut off all frequencies smaller than hmax. That is, your code does not remove a single zero frequency. In general, the amplitude at zero frequency is just an average and very often it is either not needed at all or even interferes with it.
I'm not a DSP specialist, just figured out the fft myself when I needed to, and now I wanted to help.
The matcad considers the initial series 0,1,2, 3, 4, ... The matcad considers it a real function. The fastfouriertransform considers it complex, i.e. 0+1*j, 2+3*j, ... . Perhaps norming coefficients are also taken into account differently, I'm not using matcad myself and cannot say for sure.
I still managed to make an addition to the previous post
The realfastfouriertransform also fails to pass the imaginary part of the first number, and the normalisation is not clear at all. Without understanding of this question there is no sense to calculate energy,
2007.11.03 00:36:49 2007.10.01 00:00 proverka GBPUSD,H1: i=6 Input=6 Output aa[i]=-1.1716; aa[i*2]=0; aa[i*2+1]=0
2007.11.03 00:36:49 2007.10.01 00:00 proverka GBPUSD,H1: i=5 Entry=5 Exit aa[i]=3; aa[i*2]=0; aa[i*2+1]=0
2007.11.03 00:36:49 2007.10.01 00:00 proverka GBPUSD,H1: i=4 Entry=4 Exit aa[i]=4; aa[i*2]=0; aa[i*2+1]=0
2007.11.03 00:36:49 2007.10.01 00:00 proverka GBPUSD,H1: i=3 Entry=3 Exit aa[i]=-6.8284; aa[i*2]=-1. 1716; aa[i*2+1]=0
2007.11.03 00:36:49 2007.10.01 00:00 proverka GBPUSD,H1: i=2 Entry=2 Exit aa[i]=-6.8284; aa[i*2]=-4; aa[i*2+1]=3
2007.11.03 00:36:49 2007.10.01 00:00 proverka GBPUSD,H1: i=1 Entry=1 Exit aa[i]=3; aa[i*2]=-6. 8284; aa[i*2+1]=-6.8284
2007.11.03 00:36:49 2007.10.01 00:00 GBPUSD,H1: i=0 Entry=0 Exit aa[i]=21; aa[i*2]=21; aa[i*2+1]=3
realfastfouriertransform also does not pass there is imaginary part at the first number, and normalization is not clear at all.
The first number has no imaginary part, so under index 1 realfastfouriertransform writes frequency amplitude N/2, which also has no imaginary part. By the way, it is clearly spelled out in my indicator. And here is a familiar picture from the known address
P.S. Normalization in this case is a constant, that is, if you do not take it into account any ratios will not be violated, it is the same as the measurement in centimeters instead of meters.
Thank you, I was not paying attention. Unfortunately I'm not so fluent in MQL yet to find this processing in your indicator without any comments. I have to deal with rationing now.
Edit
Yeah I found it, just multiply by n. I deleted my post to the library, it is correct.
Victor(Vinin), why did you delete your topic (" Market Phase Analysis")? It was a good topic, there wasn't any swearing either...