Neural networks, how to master them, where to start? - page 9
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
This is correct: biparametric exponential smoothing is NOT inferior to dual-input NS.
It is inappropriate to compare exponential smoothing and NS because different mathematical apparatuses are used.
different apparatus.
It is inappropriate to compare exp. smoothing and NS because different mathematical apparatuses are used.
different apparatus
so prove different)))
so prove that it is different)))
As a former teacher of SPEC subjects at a university, I explain (for free):
1. Exp. smoothing adjusts the time series (usually the closing price of each bar)
Taking into account 2 or 3 parameters, if 2 parameters are taken into account, we obtain a
two-parameter exponential smoothing, and if we take into account 3 parameters, we obtain
3-parameter exponential smoothing.
1st parameter: this is the price location parameter
2nd parameter: this is the trend slope parameter
3rd parameter: This is a seasonality parameter (factor)
The first 2 parameters are calculated using the recurrent formulas:
S[n]=w*y[n]+(1-w)*(S[n-1]+T[n-1])
T[n]=t*(S[n]-S[n-1])+(1-t)*T[n-1]
then, the "predicted" value : y [n+1]=S[n]+T[n]
As initial (i.e. initial) values for the 1st and 2nd parameter we can
take the coefficients from the linear regression formula.
2. To "predict" price movement you use NS in the form of classifiers (upwards, downwards,
I don't know) - where a fundamentally different mathematical apparatus is used.
to budimir
approach to the shell score 5
Next, break down the NS with two (2=??) inputs vs 2XEMA and see what's the difference there)))
Шаг 1: Выбираем входные данные. Например,
x1 = WPR Per1
x2 = WPR Per2
x3 = WPR Per3
Am I correct in assuming that the input data means the variables in the external EA parameters to which the coefficients will be compared?
to budimir
approach to the shell score 5
Next, break down the NS with two (2=??) inputs vs 2XEMA and see what's the difference there)))
what is the difference between NS with 1000 inputs predicting HIGH and LOW and 2xEMA ?
If the optimisation tasks are the same, the difference will be
1) in NS redundancy
2) in noise of NS
if the optimization problems are different, the difference will be in the following
1. in case of 2xEMA the subsequent TC construction is added manually from some assumptions
2. However, NS itself will allegedly find and itself will allegedly confirm and allegedly implement these "assumptions" in itself, i.e. it will be supposedly sharpened for latent regularities.
= power and structure of mathematical apparatus 2хЕМА+ТС are similar to NS, i.e. a ring of operations 2хЕМА+ТС is similar to a ring of operations NS
Шаг 1: Выбираем входные данные. Например,
x1 = WPR Per1
x2 = WPR Per2
x3 = WPR Per3
Am I correct in assuming that the input data means the variables in the external parameters of the EA, with which the coefficients will be compared?
This is how I see the coefficients of a simple Expert Advisor based on fractals:
What should I do with it now?
= power and composition of mathematical apparatus of 2xEMA+TS are similar to NS, i.e. the ring of operations of 2xEMA+TS is similar to the ring of operations of NS.
so if they are similar (in terms of the mathematical apparatus), then, when choosing these methods, one can use the criterion of prices for neuro-packages and the software,
calculating 2xEMA+TS ??? - the MetaTrader itself may be suitable as the latter, i.e. the Expert Advisor and these formulas may be written in the mql-language,
Note- FREE!