Machine learning in trading: theory, models, practice and algo-trading - page 956
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
Another useful book on the subject
Good luck
Neither demo nor real money has not yet worked out for anyone.
Bottom line - it's all toys, entertainment and fun...
Neither demo nor real has worked out so far for anyone.
The result - it's all toys, amusement and fun...
Napoleon's son was given an aluminum rattle for his birthday. Aluminum was useless for anything else.)
YOOOOOOOOOOOOO!
Something both ways used the forest. In training all classes are guessed, in the test data a little less, in the test data a little more than 50% falls into the desired class and there goes the data of the negative class (in the amount of about 50% of the desired).
Well, at best there are few examples and a little less of the negative class in the search class.
That I doubt the data can be split at all, at least by a small margin that will have a meaningful effect on the trade.
Or is it not?
Intermediate results of experiments with the tree
This 2017 report - input is generated by a TC condition, with no filters applied, but with position support
the same, but the input is generated by the Tree trained on 2015 and 2016
And this is input by ATC signal with filters
And this is the input by Tree with the same filters
Yes, the filters were optimized for 2016-2017, so it's almost a prdgon, but why the tree can't line them up is a mystery. On the other hand, you can see that where the filters screened out the inputs, the tree came in and vice versa, which is just as interesting. And the interesting thing is that the tree does not take into account the exact financial result when deciding to branch, and optimization on history is aimed precisely at financial performance.
YOOOOOOOOOOOOO!
Something both ways used the forest. In training, all classes are guessed, in the test data a little less, in the test data a little more than 50% gets into the desired class and there goes the negative class data (in the amount of about 50% of the desired).
Well, at best there are few examples and a little less of the negative class in the search class.
That I doubt the data can be split at all, at least by a small margin that will have a meaningful effect on the trade.
Or does it not?
The forest is a garbage dump in Africa if you litter it with garbage, and you gave the most concrete proof that you have NO predictors at all relevant to the target variable.
A forest is a garbage dump if you fill it up with garbage, and you gave the most concrete proof that you have NO predictors at all relevant to the target variable.
Is it like they have to divide by the target variable? :DDDD
show me a test and trace graph where the relation is present
i don't think he knows what he's writing about... he certainly didn't write about searching for predictors through the forest and other packages, since the ratio can't be found in the market using such methods, it's still a fitting anyway
A forest is a garbage dump if you fill it with garbage, and you gave the most concrete proof that you have NO predictors at all relevant to the target variable.
Here you go. Lots of examples show when train=>validation. And you need train=>validation=>test(test data, which the algorithm doesn't see at all, but only predicts by trained model, on train, validation)
So those examples that show train results and then on validation don't say anything. I have plenty of examples where validation succeeds in guessing 95% of targets.
And they use up to k-10 cross-check. I still get overtraining.
I have plenty of examples where you can guess 95% of the targets.
found, found where I wrote:
M. Gunther. Axioms of a Stock Speculator:
Auxiliary Axiom #5. Beware of the trap of historical parallels.
Auxiliary axiom No. 6. Beware of the illusion of repeating figures.
Auxiliary axiom #7. Beware of the delusion that correlation and causation exist.
hmm, not a bad match to Gunter? ))))
found it, found where I wrote:
Hmmm, not a bad match to Gunther? ))))
I've known this for a long, long time. I think I went through it in "third grade." Not surprised))))