Machine learning in trading: theory, models, practice and algo-trading - page 1521
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
Wow, even in Russian, thank you.
We started with good health and ended with Elliott waves. No mathematical methods for predicting financial times are offered, except for determining the persistence of the series, which is already clear.
This is just muddled.
We began in good health and ended with Elliott waves. No mathematical methods for predicting financial times are offered, except for determining the persistence of the series, which is already clear.
What a mess.
So, again nothing new, eh.
So I'll keep on splitting the tree.
The graph iterates splitting with their leaves (X) in the form of profit accumulation on each leaf (Y) of the split outside the training sample.
It follows from the graph that splitting the tree can give alternative versions of logical chains, including those describing a different area of space, but with the same or better results.
The problem here is that all leaves need to be combined into one system and given weights based on their correlation between each other and completeness of description of the sample space, and I don't know what formulas to operate with in this case, unfortunately.
Well, that means nothing new again, eh.
So I will continue to split the tree.
The graph iterates splitting with their leaves (X) in the form of profit accumulation on each leaf (Y) of the split outside the training sample.
It follows from the graph that splitting the tree can give alternative versions of logical chains, including those describing a different area of space, but with the same or better results.
The problem here is that all the leaves need to be gathered into one system and given weights based on their correlation between each other and the completeness of the description of the sample space, and I do not know what formulas to operate with in this case, unfortunately.
I noticed that if you divide the market into conditional states, train the model for each of them and switch between them, the results are better
i'm working on hmm for that
I noticed that if you divide the market into conditional states, train the model for each state and switch between them, the results are better
That's what I'm doing with hmm.
It is logical to be better but how to distinguish the states?
It makes sense that it is better, but how to distinguish between states?
Either on automata (clustering), without a teacher. I.e. several hidden states will be distinguished, which will obviously differ by stat. stats. Examples were thrown above.
or at random
Either on automata (clustering), without a teacher. I.e. several hidden states will be distinguished, which will obviously differ by stat. stats. Examples were thrown above.
Or at random.
Any clustering method found for this? After all, it should have, shall we say, smooth transitions from cluster to cluster, if it identifies global trends, there should be a kind of sensitivity and logic of transition from one state to another.
Has any clustering method been found for this? After all, it should have, shall we say, smooth transitions from cluster to cluster, if it identifies global trends, there should be a kind of sensitivity and logic of transition from one state to another.
they already wrote about HMM
The guys on the ladder against the dipmind (accidentally caught by them)
People on the ladder against dipmind roll (accidentally caught them)
I also read about AI in games, but a lot of non-NS logic in game bots, often using decision trees.
I've found Rete algorithm, I don't know why, but I've never heard of it, I've found this reading, but it's too big - 1000 pages, something in it...
I also read about AI in games, but a lot of non-NS logic in game bots, often use decision trees.
I came across Rete algorithm, I don't know why, but I've never heard of Rete algorithm, I found a reading, but it's a very big volume of 1000 pages, there's something in it...
Well, they don't do reinforcements on trees...
If i play a lot of games, how do i retrain them?