Zero sample correlation does not necessarily mean there is no linear relationship - page 7
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 don't understand.
So here we have the USDJPY. The range is 83.15 to 85.9.
And the euras range is 1.31 to 1.37.
How do we convert the USDJPY to the EUR range?
USDJPY ' = EURUSD.min + (Var - USDJPY.min) / (USDJPY.max - USDJPY.min) * (EURUSD.max - EURUSD.min)
.
With linear regression and RMS the normalization seems to be correct. (?)
So here we have the USDJPY. The range is 83.15 to 85.9.
And the euras range is 1.31 to 1.37.
How do we convert the USDJPY to the EUR range?
USDJPY ' = EURUSD.min + (Var - USDJPY.min) / (USDJPY.max - USDJPY.min) * (EURUSD.max - EURUSD.min)
Theoretically, it is possible. In practice it is practically suicide: to search for min and max on each window and transformations along the length of the entire sample each time. Wouldn't it be easier to simply prolagarithmize ONE time?
With linear regression and RMS I think I wrote the normalization correctly. (?)
Theoretically, it is possible to do so. Practically, it's almost suicidal to look for a minimum and maximum on each window and do a transformation every time along the length of the whole sample. Wouldn't it be easier to just ONE time prolagarithm?
What makes you think that linear regression is defined by max and min? And what is the practical effect of this?
This is the second way. Linear regression is y = kx + b, you find the coefficients k and b.
The question is - why is it worse than logarithm?
.
P.S.: let's say there are three ways of normalisation. How to quantify which is better ;-) ?
The car will count. His head is iron :-).
This is the second way. Linear regression is y = kx + b, you find the coefficients k, b.
The question here is: how is it worse than logarithm-wise?
The regression is searched by MNC, not the way you wrote.
The regression is looked for by the ISC, not as you wrote.
I listed two ways.
The first is with min and max.
The second is a linear reg. I didn't write anywhere about how to calculate it.
I didn't write how to look for regression.
I listed two ways.
The first is with min and max.
The second is a linear reg. I didn't write anywhere about how to calculate it.
I thought you wrote equivalents of the same thing.
The regression option is wrong. The conversion option is better, but also bad.
What's the point? Tell me the development methodology-and I'll tell you what you get.
If Mq4-indicator matched Mathcad, what could be the point of argument?
The fact that the indicator showed the same thing is a clear diagnosis. "Healthy".
.
If you can, please write what you think about the calculation that hrenfx is talking about.
When two offset windows are taken and the line and RMS are counted in them separately - and the corr.
The method is naive, but somehow it evokes sympathy.)
As for what hrenfx says (if I understand correctly) it may be called in trader's terms search for patterns on history. Set of ready windows of the history (patterns) is compared to the current one. If it coincides, then we kind of know what to do, if we assume that history repeats itself...
As for what hrenfx says (if I understand correctly of course) it can be called in trader terms. searching for patterns on history.
It's about calculating the sample characteristics of BP.
This is obvious :-) - Get a graph - and "plug in" on it.
Construct the expectation of autocorrelation.
.
The idea about patterns is on a different plane.
So these two ideas do not contradict each other.
Construct the expectation of autocorrelation.