neural network and inputs - page 12
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
In general, I will tell you this - I personally (3 times imho) sometimes use neural networks, as I wrote above, but as an intermediate stage, so to speak, but in my memory nets have not helped me anything really (while other methods gave results). I don't know why, maybe I just use them "in the first preparation" so to speak, without going deeper and not really bothering with the architecture. I.e., as I wrote above, if I have, for example, an idea, I go at it both ways, including with nets, just for order, so that in case of something "not to spill the baby". However, as I said above, not once have they given me anything tangible. At least compared to the range of other methods (you know, regressors, solvers, nearest neighbours, etc.).
I think the main problem is overtraining (the antithesis of robustness). Actually we all are after robustness, whether we realize it or not. Things are very bad in grids - worse than in other algorithms. According to my observations, the simpler a solver is, the more chances it will be robust.
You say everything correctly, but there is a very big and thick BUT:
If there is no pattern in the data, but there are random coincidences in the analysed section, the NS will find them and give them out as a pattern.
If in some part of the trajectory of a ballistic missile with a nuclear warhead, its path coincided with the route of seasonal migration of cranes, then this does not mean that we can predict the entire route of the missile following a known migration path. But the NS will output the result as a revealed pattern.
it depends on how the network is set up
it depends on how the network is set up
If there is no pattern in the data, but there are random coincidences in the analyzed area, then the NS will find them and present them as a pattern.
If a ballistic missile with a nuclear warhead had its flight path coinciding with the seasonal migration route of the cranes, it does not mean that one can predict the entire flight path of the missile using the known migration path. But the NS will output the result as a revealed pattern.
Well, that's the skill of a trader - to choose such a segment for analysis, which would characterize the market in the future. In this case the neuronet will trade successfully (profitably) in the future, and not in the past that has been discussed many times on this forum ))).
...to select a section for analysis that would characterise the market in the future...
Are you confusing traders with fortune tellers, by any chance? :) How do you know how it will be in the future? The future is uncertain (c)
Are you confusing traders with fortune-tellers, by any chance? :) How do you know what the future holds? The future is uncertain (c)
...Don't read anything...
Any TS, with or without neuronetworks, uses patterns that are searched on past data. So in essence there is no guarantee of making money on these found patterns in the future. Or do you have some method of determining that the patterns found on past data can be used to make money in the future?
I've written before...
But from a mobile phone...
I suggest to look at PAMM Wetteg on mcl 5. He is an expert in nerve-racking and the winner of one of the championships.....