Machine learning in trading: theory, models, practice and algo-trading - page 1805
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
I do not want to bother, synchronize with the data tester ... Bae :) although ... what the hell
news in the tester is the best havI couldn't find another one, and I downloaded the file about 4 months ago))))
It's clear.
If you can determine the durability time of the TS - if you can, then everything is OK - traded its time - taught - traded - taught ...
I don't have enough data for the test, that's why I suggest checking both the left and the right side, while it makes no sense to look deep into the history, volatility changes all the time, that's why only a couple of years are relevant, at least I use genetics for 1.5 year options, I look for forwardtest for 6 months, sometimes for a year, more than a year the TS appears with big drawdown - not the one I would like
I have no qualitative signs, I have to make up a bicycle. If the 5X is rejected from the trane, it is a good model anyway.
Is there any way to find out when the training parameters change? Something in between is needed. If there is one algorithm, it is the same one, and you probably need some intermediate results.
Is there any way to know when the learning parameters change?
I don't know, everything is automatic. If I look at the ratio of profitable/lossmaking trades, the total revision can tell what has started to work worse
I have everything on automatic. According to the ratio of profitable/lossmaking trades, the total revord can be determined that started to work worse
Late as always is a big deal. Some crutches are necessary. If I wanted to use different mathematical models and see which one is better.
I have everything on automatic. According to the ratio of profitable/lossmaking trades, the total revord can be determined that it started to work worse.
The learning period should not be increased but decreased.
The lateness is as great as always. Crutches are needed. Maybe check different matrices and see when which is better?
well it's state space models, works every once in a while too
Well, these are state space models, they work in some cases, too.
Well, we have a minimum time task to determine what is wrong with the series) Initially it is assumed that the stationary series with moving average described by the mathematical model will give sufficient results at MO. When the parameters of the mathematical model are changed, there is nothing wrong and the period before learning is acceptable. When we break / change the model, we have a left plot that is not correct for the new model, and we do not know exactly the required period for learning.
We need something inside training, like an indexer.
Well, we have a minimum time task to determine what is wrong with the series) Initially it is assumed that the stationary series with the moving average described by the mathematical model will give sufficient results in the case of MO. When the parameters of the matrix model are changed, there is nothing wrong and the period before learning is acceptable. When we break / change the model, we have a left plot that is not correct for the new model, and we do not know exactly the required period for learning.
Something inside training is needed, like an indexer.
I have made several models depending on the indices values, each model for a specific range of indices values. I have always tried to use some draws on them.
made several models, depending on the values of the indices, each model for a particular range of values of the induced. Sometimes it helps, sometimes not.