Predicting the future with Fourier transforms - page 11
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
To all : Should we use the indicator on pairs with as little news as possible, which will throw off the analysis? You can't predict them.
ANG3110: Judging by the picture it's a very good script:) I think I want to play. Knock on the chat when you have some free time? 178971187
As experience shows news points just coincide well with the forecast. The quality of predictions is more influenced by the ability to adjust and assess the situation over several periods. I don't think the timeframe is very important for the script, it's more important where there are good wave resonances, and they are often quite visible if you walk a bit through the recent history and identify the most frequently occurring periods and then try to make a prediction for the current day or two.
I've typed in my ICQ, but you're not online, and I turn mine on by appointment via email. So if you want to talk to me in advance, the address is in the post above.
Experience has shown that news points are just as good as the forecast.
What is a TS? It is a thechanalysis. Experience has shown that a properly done theanalysis coincides with news movements 70% of the time.....
What is TC? It's a thechanalysis. Experience shows that a properly done thechanalysis coincides with news movements in 70% of cases.....
Well probably the Neural Networks fans, and in particular NeuroShell (for which I have the utmost respect for their thoughtful presentation), probably understand this more than anyone else.
To all : Should we use the indicator on pairs with as little news as possible, which will throw off the analysis? You can't predict them.
ANG3110: Judging by the picture it's a very good script:) I think I want to play. Knock on the chat when you have some free time? 178971187
As experience shows news points just coincide well with the forecast. The quality of predictions is more influenced by the ability to adjust and assess the situation over several periods. I don't think the timeframe is very important for the script, it's more important where there are good wave resonances, and they are often observed quite well if you walk a bit through the recent history and identify the most frequently occurring periods and then try to make a prediction for the current day or two.
I've typed in my ICQ, but you're not online, and I turn mine on by appointment via email. So, if you want to talk to me, just write me the address above.
What is TC? It is a tehanalysis. As experience shows, properly performed thechanalysis in 70% of cases coincides with movements on the news.....
Well probably fans of Neural Networks, and in particular NeuroShell (for which I have the utmost respect for their thoughtful presentation), probably no one else understands this.
I won't argue with experience, of course, but it's far from obvious that news has to fit the forecast. The same neural networks should only be used when the market is calm, i.e. when it is not shaken by the misunderstood words of some finance minister. We are looking for patterns - by neural network or Fourier, or even more complicated - and who is going to find a pattern in this minister's intonation? I doubt even he himself could do it.
Anyway, I'll write a letter, all right.
Hello!
Can this indicator show only Value5 (FFT and Future Search v2.1), and do not do any other calculations. it takes a lot of resources - actually loads the whole processor. especially if enabled on several instruments.
Thanks in advance!
The topic is very interesting.
Predicting the future with Fourier is of course possible ... for (quasi) periodic processes. And since there are always some waves in the market, something may work out.
To identify periodic components, we can do the following:
do a Fourier analysis on the minutes with a sample width of a week (for example) and see if there are harmonics expressed there. Then, move the weekly window by minutes and see how the amplitude/frequency of the harmonics (spectrum peaks) change as a result of that shift.
If there are no expressed harmonics, it seems that Fourier will not work.
The topic is very interesting.
Predicting the future with Fourier is of course possible ... for (quasi) periodic processes. And since there are always some waves in the market, something may work out.
To identify periodic components, we can do the following:
do a Fourier analysis on the minutes with a sample width of a week (for example) and see if there are harmonics expressed there. Then, move the weekly window by minutes and see how the amplitude/frequency of the harmonics (spectrum peaks) change as a result of that shift.
If there are no expressed harmonics, Fourier probably won't work.
Why don't we predict not the price, but a price indicator, MACD type, for sure, there are waves there
Or maybe not the price, but a price indicator like MACD, that's where the waves are for sure
In general, the best way is to "predict" profit. There is no point in predicting any indicator because it has many false entries.
there are such indicators on the Fourier theme :
satop, beautiful )
If it's not hard to forecast where the Jap will go by tonight.