Stochastic resonance - page 24
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
I propose to use DSP for noise calculation and analysis, to clarify my thought I will give a brief example in MathCad
ANALOG SIGNAL FILTERING
USING BPF
A two component complex signal q is synthesized below
Signal s is signal q with a noise component superimposed on it
Using a direct FFT, the time domain representation of signal s
is converted to the frequency domain, harmonics with small amplitude of signal (smaller a) are filtered out and then inverse FFT ensures return to the time representation of the signal. The degree of filtering is determined by the value of parameter a.
What this gives us:
The problem is that the analyzed data contains both signal and noise. In practice, to measure noise in radars (radar stations) close the receiver and measure the noise. It is used to set thresholds. When receiver is opened, incoming data (signal+noise) is analyzed and if threshold is exceeded, decision is made on signal detection.
By the way from a mathematical point of view the best performance has the Waldo detector (two thresholds)
I'm sorry, very busy with 3 jobs + trying to do forex in the rest of the time. Would love to join the working group. I think the best way is to do a Skype chat. If anyone is interested, knock privalov-sv. I have many ideas and they have been tested in practice not related to the forex market. It would be good to check them. But my hands, and most importantly the time is not enough.
P.S. If you can teach me how to get the minute quotes from 1999 in text format. I only have 65000 and I cannot get them. Thanks in advance.
Minute quotations from 1999 can be downloaded from History Center. You can watch a short video in the thread How works downloading from History Center
How to download, I know :), I did it a long time ago, but how to convert them ALL into a text format, what would be convenient for analyzing in Matkadec?
...
Minute quotes from 1999 can be downloaded from History Center. You can see a short video in the branch How works downloading from History Center
How to download, I know :), I did it a long time ago, but how to convert them ALL into a text format, what would be convenient for analyzing in Matkadec?
The proposed method is very poor, with strong edge effects, as you can see in the graphs, it is not actually used in practice and cannot be used for this task. It is difficult to call what appears at the edges noise, and for real time noise intensity control it is by no means suitable.
Please tell me how to scientifically calculate noise intensity, because we had a lot of options here.
There is an export button in "chistorcenter", press it and you will get a text file.
Minute quotations from 1999 can be downloaded from History Center. You can watch a short video in the thread How works downloading from History Center
How to download, I know :), I did it a long time ago, but how to convert them ALL into a text format, what would be convenient for analyzing in Matkadec?
hst2csv.mq4 - MQL4 Code Base
https://www.mql5.com/zh/code/7719
Minute quotes from 1999 can be downloaded from History Center. You can watch a short video in the thread How works downloading from History Center
How to download them I know :), I did it long ago, but how to convert them ALL into a text format, what would be convenient for analyzing in Matcad?
Hi Seryoga!!!
Good to see you! What do you say about stochastic resonance? How do you scientifically calculate noise intensity?
Hi Seryoga!!!
Good to see you! What do you say about stochastic resonance? How to scientifically calculate the intensity of the noise?
Hi Sergei!!!
Likewise!!!
As for intensity, scientifically, it has always been the square of the amplitude with some dimensional constant.
I.e. you have to centre the noise - subtract the constant component (if any) and take the integral (sum) of the amplitude squared by the integration window at the time interval of interest.
That is all.
P.S. I did nothing but make an interesting indicator for MT4. It shows attraction of a price series to a linear or parabolic trend (k=0 - flat, k=1 - linear trend, k=2 - parabolic, n - sampling limit).
The red window in the figure shows the polynomial coefficients using the least squares method and the forecasting window in blue. You can really observe the effect of attraction of time series to equilibrium state (or vice versa :-)!
Hi Seryoga!!!
Good to see you! What do you say on stochastic resonance? How do you scientifically calculate noise intensity?
Hi Sergei!!!
Likewise!!!
As for intensity, scientifically, it has always been the square of the amplitude with some dimensional constant.
I.e. you have to centre the noise - subtract the constant component (if any) and take the integral (sum) of the amplitude squared by the integration window at the time interval of interest.
That is all.
P.S. I did nothing but make an interesting indicator for MT4. It shows attraction of a price series to a linear or parabolic trend (k=0 - flat, k=1 - linear trend, k=2 - parabolic, n - it is sampling).
...
The red window in the figure shows the least-squares window for the polynomial coefficients and the blue one for the prediction window. You can actually observe the effect of time series attraction to the equilibrium state (or vice versa:-)!
Thanks, I'll try some more scientific calculation of noise intensity. Although the square of the amplitude reminds me a bit of signal power, but it doesn't matter.
Cool, the indicator should probably be of interest to Candida. Surely she will see the intrigues of potential power :o)))
I see people are getting more and more interested :)
Cool, the indicator must surely be of interest to Candida.
Well, it's more of a tool for the doubters - it gives you an opportunity to twist and estimate. But it definitely demonstrates the business mindset :)