Machine learning in trading: theory, models, practice and algo-trading - page 2265
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
The same shit with paired trading, you're looking for a nice spread, but on the OOS it immediately disperses.
What to do?
The same crap with paired trading, you look for a nice spread, and on the oos it immediately disperses.
eva and chif, 10 days
they already explained above
MO is the coolest fit, a tester can't do that
Guys, explain to the nubus, what does the MO and forex have to do with it, where is the connection?
They have already explained above
The Expert Advisor is the best fit, a tester cannot do that.
Forum on trading, automated trading systems and testing trading strategies
Machine Learning in Trading: Theory, Practice, Trading and Beyond
mytarmailS, 2020.12.24 11:08
I don't know how to use automatic selection of parameters in the Strategy Tester, this is already a primitive machine learning.
eva and chif, 10 days.
And before that, it was spreading. If you build a spread on history from 7 to 16, after 16 the spread will split.
Before that, it was spreading. If you build a spread on history from 7 to 16, after 16 the spread will spread out.
ok
let's see
And I slowly continue to experiment with training neural networks with fitness function ....
I came up with this way of describing the utility function (fitness function) , instead of teaching the network to maximize profit, I tried to teach the network to "maximize beautiful graph of income"
what is "maximally beautiful income graph" ? i expressed it as coefficient of correlation of linearly increasing line and yield graph
traine 20k
test 20k
the criterion - "the most beautiful graph of income " (described above) + double commission
the neuron is trained poorly, too few neurons for such a large sample
but the criterion clearly describes the model better by equity than max profit
Before that, it was spreading. If you build a spread on the history from 7 to 16, after 16 the spread will spread out.
Thank you! You made me laugh with your three posts, from the bottom of my heart!
Especially touched my soul.
"And then the fantasy concile...
"Darkness. Darkness. The factory."
Train 20k
test 20k
criterion - "the most beautiful graph of income" (described above) + double commission
neuronics is poorly trained, too few neurons for such a large sample
but the criterion clearly describes the model better by equity than max profit
Proposal - a traine of 20k split into three more sections and
criterion -"maximally beautiful income graph" (described above) + double commission - to supplement it + maximally similar at all three segments
I.e., not only that, the average "beautiful" of the three plots should be better, and the difference between the minimum and maximum "beautiful" of the three plots should be minimized.
Train 20k
test 20k
criterion - "the most beautiful graph of income" (described above) + double commission
neuronics is trained poorly, too few neurons for such a large sample
but the criterion clearly describes the model better by equity than max profit
it's ok, go to Market with a trillion monthly turnover