Machine learning in trading: theory, models, practice and algo-trading - page 3056
![MQL5 - Language of trade strategies built-in the MetaTrader 5 client terminal](https://c.mql5.com/i/registerlandings/logo-2.png)
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
Anything on the subject? That's funny.
Isn't that on topic?
What's so funny about it?At the matstat level I guess. If on average several models are wrong in predicting the same thing on new data (on validation subsample), then it is unpredictable at all and is moved to "do not trade"
You can take a specific time and corresponding values of signs/signals as the "same thing"Sometimes did a re-fitting (with rubbish thrown out), getting a quilt with holes. Then classified the remaining quilt pieces and darned some holes, getting already small quilts - islands of regular shape where there is a pattern. After that, trained on each small quilt without throwing anything away anymore.
In this way, overfitting helped to quickly identify islands of predictability. At the same time to get away from re-fitting.
For example, I found long-lasting patterns lasting 30 minutes in the afternoon.
Sometimes did a re-fitting (with trash thrown in), getting a quilt with holes. Then I classified the remaining quilt pieces and darned some of the holes, resulting in small quilts - islands of regular shape, where there is a pattern. After that, trained on each small quilt without throwing anything away.
In this way, overfitting helped to quickly identify islands of predictability. At the same time, it helped to get away from overfitting.
For example, I found long-lasting patterns lasting 30 minutes in the afternoon.
Has the hairy ball rolled around on small blankets yet? :)
Yes. Otherwise, the point slips away.
Yeah. Otherwise, the point slips away.
It can also be expressed through bread when comparing subsamples without and with training by means of ols regression coefficients (estimation of the effect of "treatment").
T=1 samples with treatment, T=0 without, mean is the difference of whether there was a treatment effect on average
I'm still a noble at causal inference.
It is also possible to express through bread when comparing subsamples without and with training by means of ols regression coefficients (estimation of the "treatment" effect).
T=1 samples with treatment, T=0 without, mean is the difference whether there was a treatment effect on average
I have a weakness with associations. I have zero understanding of ME.
I'm still a sucker for causal inference.
It's just a bad format of communication and you're weird. Definitely a good incentive to think deeply - really getting kicked in the teeth by the market.
the graphs are generated by a random function
Is it possible to distinguish from the real ones????
all candlestick configurations, eskimo, takeovers... it's all there.
What's real, what's illusion of the mind?
And my author's technique of precise inputs works there too, ON RANDOM!!! how is it even possible?????
You can model all kinds of trends and different situations, then calculate the parameters of the TS
"Dressed in candlesticks" your sinusoidal model
What is a reversal in terms of candlesticks on the model of two sinusoids?
It's when volatility drops to statistical lows in major waves.
You can see what a trend entry is here.
graphs are generated by a random function
Is it possible to distinguish from real ones????
all candlestick configurations, eskimo, takeovers... it's all there.
What is reality, what is illusion of the mind...
And my author's technique of precise inputs also works there, ON RANDOM!!! how is it even possible?????
Some code that doesn't work for me again. If you want substantive discussions, post reproducible results.