Neural networks. Questions from the experts. - page 2
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
Then another question )))) Is there a correlation between the minimum error on the test one and the maximum profit?
Then another question )))) Is there a correlation between minimum error on a test and maximum profit?
There is, but there is no formula (universal as in mathematics), the relationship for each strategy/task has to be deduced.
There is, but there is no formula (universal, as in mathematics), the relationship for each strategy/task has to be deduced.
Well, I, for example, in my experience did not feel that minimum error on OOS unconditionally leads to maximum profit.
No, minimum error itself does not mean maximum profit, what I meant was that the link between error and profit can be established.
For example:
error 0.6 - (-1000 p)
0.55 error - 0 p
error 0.5 - 1500 p
error 0.45p - 3800 points
No, minimum error itself does not mean maximum profit, what I meant was that the link between error and profit can be established.
For example:
error 0.6 - (-1000 p)
0.55 error - 0 p
error 0.5 - 1500 p
error 0.45p - 3800 points
I agree, but we are fighting for the maximum profit. And here, the minimum error does not give us the maximum profit. Well at least I couldn't find any confirmation of this with myself......
Agreed, but we are fighting for maximum profits...
And the non-maximum, but smooth profit growth is no longer satisfactory (:
Or maximum profit is so small, that it barely covers spread?
And the non-maximum, but smooth profit growth is no longer satisfactory (:
Or is the maximum profit so small that it barely covers the spread?
OK - maximum profit, which is achieved by a smooth growth of equity, without significant drawdowns. )))
Then another question )))) Is there a correlation between the minimum error on the test one and the maximum profit?
As far as I understand, the author didn't have such a task %)
1) Did I understand correctly that a neural network is not able to reconstruct a function if it is inherently dynamic as in the case of ACC, even having all the necessary data to calculate it, since if the formula is rigidly static as in the case of LVSS or EMA, there is no problem.
2) If I am wrong, which networks should be used? And used MLP in statistics.
3) I have heard the opinion that auto nets and nets of own e.... design, if I may say so, there is not fundamentally much difference. Is this really the case?
4) What networks and what programs would you advise for application to financial markets, in particular for the task I have described, i.e. to restore values from all known data.
Point 1 According to the paradigm multilayer NS is capable of restoring any function, the question in practice most often comes down to data preparation and training methodology. Alternatively, you could try varying the training data. By the way, since in adaptive mean volatility varies over time, I suggest you try the sliding window method to form a training sample.
Point 2 MLP is enough, there are quite a few different NS architectures based on it.
p.3 Well, if it's implemented correctly, what difference does it make!
p.4 Matlab, as an architecture I'll suggest any variant of recurrent network, though MLP should suffice ...