Machine learning in trading: theory, models, practice and algo-trading - page 2340

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
Something is getting in the way.
There is no understanding of what to do. And the network, even the most cunning - itself dumb, can not guess. And when understanding does appear - the network is no longer needed. Paradox of Nature :)
There is no understanding of what to do. And the network, even the most cunning - itself dumb, can not figure it out. And when understanding does appear - the network is no longer needed. Paradox of Nature :)
If you stop repeating like parrots the same thing and start doing something, you can stop looking like a specimen of a dead-end branch of evolution
Who knows about processor operations?
I converted the quick sorting function from double to float type. I just changed the data type.
It would seem that it should be 2 times faster with 4 bytes than with 8 bytes. But during metering the gain is only 5-10%.
The main operations are comparing numbers and rearranging them in memory.
Rearrangement should be faster, because 4 bytes instead of 8 bytes need to be rewritten. Or does the processor spend the same time on this?
Comparison probably isn't accelerated because we have all data with 5 decimal places, while comparison operation most likely goes bit by bit, as soon as it gets to the bit where 1 number is bigger than another, it finishes the search of number bits. I.e., for both double and float, the enumeration of bits will stop after 10-20 bits, on the same bit.
Do I understand it correctly? Or is there any other explanation?
No, of course not. The compare, add, subtract, write, read, copy functions themselves are long, and not at all linearly dependent on data type.
And if you stop repeating like parrots the same thing and start doing something, you can stop looking like a specimen of a dead-end branch of evolution
That's for evolution to decide. Do you even have a girlfriend? Or is there no lamba? :))))
Evolution will decide that. Do you even have a girlfriend? Or is there no one without a lamb? :))))
Just offered you a way out of the situation, troll. You can just Up the wall, a paradox of nature.
https://www.mql5.com/ru/forum/86386/page2317#comment_20471596
It turns out about the same (five minutes). The model breaks +- at the same time. Train for the last two months.
your algorithm pulls out the patterns in a funny way
https://www.mql5.com/ru/forum/86386/page2317#comment_20471596
It turns out about the same (five minutes). The model breaks +- at the same time. Train last 2 months.
Can you get the drain points in the data?
https://www.mql5.com/ru/forum/86386/page2317#comment_20471596
It turns out about the same (five minutes). The model breaks +- at the same time. Train for the last two months.
Funny how your algorithm pulls patterns.
If I train with 9 setups like yours, it "drags" a bit longer.
Can you get the drainage points in the data?
probably
probably
they are kind of obvious. The sad thing is that it is not clear what is needed to understand how the drain states differ from the normal state. I would work around them. In terms of a filter. )