Machine learning in trading: theory, models, practice and algo-trading - page 3681
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 apparently the pattern changes over time. Try to train on a smaller number of strings (with less deep learning) or on the contrary on a larger number of strings (with deeper learning).
I'll try it, at least that's the easiest thing to do.
Interesting.
The idea that Dipsic is first of all a careful, thoughtful work with a fitness function is confirmed. Apparently the developers' experience in applying MO to trading has had an impact.
The idea of a DeepSeek R1 + Browser Use bundle is also interesting - more on that in a bit. I hope for a soon appearance of an article on the forum about AI-imitation of manual trading via web terminal).
Purely theoretical question. I'm thinking about applying ensemble ideas from MO to a set of TCs.
The simplest, probably, is bagging - just a portfolio of independent TSs. The total position is defined as the sum or average of the positions of the individual TCs.
Staking - a system of several interacting TSs, where the work of some TSs somehow depends on the work of others. The idea is more or less clear, although it is difficult to describe it unambiguously - there can be a huge number of variants.
It is not clear with bousting. Perhaps something like a sequence of TSs, each of which improves the performance of a portfolio of previous TSs (reduces drawdown, etc.).
Purely theoretical question. I am thinking about the application of ensemble ideas from MO to a set of TCs.
The simplest, perhaps, is bagging - just a portfolio of independent TSs. The total position is defined as the sum or average of the positions of individual TSs.
Staking - a system of several interacting TSs, where the work of some TSs somehow depends on the work of others. The idea is more or less clear, although it is difficult to describe it unambiguously - there can be a huge number of variants.
It is not clear with bousting. Perhaps something like a sequence of TSs, each of which improves the performance of the portfolio of previous TSs (reduces drawdown, etc.).
Imha, all these approaches are more related to curvafitting than to the inference of any statistical regularities. In general, the whole IO paradigm is seen more as building models for prediction in case of iid, but not for causal inference.
Imho, curvafitting is only bad if the "curve" does not exist in reality. If the existence of the pattern is more or less certain, then curvafitting is an inevitable step for its practical application.
Of course, for trading, the very existence of a pattern is always the most important issue.
Imho, curvafitting is bad only if the "curve" does not exist in reality. If the existence of the pattern is more or less certain, then curvafitting is an inevitable step for its practical application.
Of course, for trading, the very existence of a pattern is always the most important issue.
Yes, but it is simply easier to find any "regularities" by any other simple method, and then train the model to predict them. In this case, the efficiency of bousting or bagging will not be much different :)
Purely theoretical question. I am thinking about the application of ensemble ideas from MO to a set of TCs.
The simplest, perhaps, is bagging - just a portfolio of independent TSs. The total position is defined as the sum or average of the positions of individual TSs.
Staking - a system of several interacting TSs, where the work of some TSs somehow depends on the work of others. The idea is more or less clear, although it is difficult to describe it unambiguously - there can be a huge number of variants.
It is not clear with bousting. Perhaps something like a sequence of TSs, each of which improves the performance of the portfolio of previous TSs (reduces drawdown, etc.).
Dipsic doesn't seem to have had much of an impact on the IT giants' plans to invest a ton of money in AI.
Maybe it's time to buy NVIDIA stock)