What 4 factors do you think the price depends on? - page 8

New comment
 
mmmoguschiy-new:
It's all more simply called foundation.
AFoundation is a fuzzy concept, I named specific numbers that can be looked at and are published. Smaller fluctuations in price depend on short-term trading, but all speculation is short-term and real demand is long-term.
 
Yousufkhodja Sultonov:

Here, as an example, is the processing of the weekly data from 02 to 06 May 2016:


The equation turns out Euro = 1.79847 + 0.000413*$ - 0.00039*gold - 0.83114*franc +0.422074*pound. We see a complete coincidence of the actual and calculated euro/dollar prices, as there are only 5 initial data points. It could not be otherwise, as this procedure is a kind of test for impeccability of formulas and algorithm. When the amount of data is larger, each tool will show its true face. If you try to recalculate on a calculator, you will not get a match, because not all significant figures are given, which are in the bowels of the computer. I guess it is understandable. I show those who doubt that the calculation formulas and the algorithm itself are impeccable. Now, the algorithm can handle any amount of data and any tool or factor.

Have read the thread, the idea of finding coefficients is a good one, I worked on it a couple of years ago. It is possible to find odds, but the problem is that they are constantly changing. The basic pattern of the market is that it is constantly changing.
 
If only we could find an algorithm to change these coefficients!
 
Maxim Romanov:
I wish I could find an algorithm to change these coefficients!
If the coefficients change, then the wrong factors have been taken and/or the dependence is non-linear.
 
You take an ordinary neural network, MLP type, input your 4 or more factors and output the price, train it and then check the results on the control section. I trained more than one layer of neurons )) result, in most cases, will be nothing... but maybe someday you'll get lucky ) And yes, you will have to wait a long time.
 
Дмитрий:
If the coefficients change, then the wrong factors have been taken and/or the dependence is non-linear.
It's like hackers discussing...
 
Дмитрий:
If the coefficients change, then the wrong factors have been taken and/or the dependence is non-linear.
The coefficients will change in either case -- (whether the factors are the same or different) (linear or non-linear) -- there will be changes.
 
Олег avtomat:
The coefficients will change anyway.

another nonsense.

If the model is adequate for a stationary process - the coefficients will not change or will change insignificantly

 
Maxim Dmitrievsky:
You take an ordinary neural network, MLP type, input your 4 or more factors and output the price, train it and then check the results on the control section. I trained more than one layer of neurons )) result, in most cases, will be nothing... but maybe someday you'll get lucky ) And yes, you'll have to wait a long time.
That's the thing, there are no stable patterns, life span of any pattern found is finite, or it will change, start to manifest itself differently. And the most difficult thing is to understand when the pattern has disappeared from the market. In the context of this topic, the coefficients will change, you can take all possible correlations and the coefficients will change.
 
Дмитрий:

another nonsense.

If the model is adequate to a stationary process - the coefficients will not change or will change insignificantly

You are talking nonsense.

You don't see the difference between the model and the input process -- get a handle on these concepts, then maybe you won't talk nonsense like this last one.

123456789101112
New comment
Reason: