Machine learning in trading: theory, models, practice and algo-trading - page 3016
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
You have a forest, do you reweight the model after building? Or do you just take the average value of the activated leaves?
That's the point, when selecting a leaf, I take into account the stability and uniformity of response distribution over history. I make two-dimensional indicators and evaluate them in aggregate. Thus, leaves without responses are an extremely rare event for me.
It seems to me that you could easily save leaves, creating thousands of trees, and work with them only.
If the average of all trees is > the required one, I use it for balance calculations.
.
How do you measure uniformity? Deviation from a straight line between the 1st and last balance points? And probably the total should be multiplied by the balance?
Oh, yeah, there's no point in using Forrest then.
What is the difference between a genetic tree and a regular tree, what are the advantages?
It differs by trying to use not the best predictor split, but different variants of the best. In this way splits are made sequentially, and the success ofevaluation is done onthe leaf, if I understand the algorithm correctly. From the successful generation, predictors closer to the leaf are cut off and the construction is retried. I can't analyse the algorithm itself in detail - I'm not the author. But, according to the idea, this approach is better than randomisation in theory.
Yes. Sometimes I use 1 tree for speed. Now I usually use several trees.
If the average of all trees > desired, I use it for balance calculations.
How do you measure uniformity? Deviation from a straight line between the 1st and last balance points? And probably the total should be multiplied by the balance?
As far as I remember, the sample is divided by years and a balance sheet is built by financial indicators, each balance sheet is evaluated by different metrics, including the topic you said, there are tolerance criteria, and if all sections (years in my case) everything is good, then the leaf is accepted into the leaf base.
As far as I remember, the sample is divided by years and a balance sheet is built by financial indicators, each balance sheet is evaluated by different metrics, including the topic that you said, there are admission criteria, and if everything is good for all plots (years in my case), then the leaf is accepted into the leaf base.
What does this have to do with 1 individual leaf? Examples in the leaf do not evenly describe the whole year, but for example 2 examples in January, 27 in February and 555 in December.
If you take the balance line from all leaves as a basis, then in December for this leaf will obviously be the main growth and the deviation from a straight line will be very strong.
If we take the balance line from only this 1 leaf as a basis, uniformity can be achieved, but the participation in the overall uniformity is difficult to determine.
What does 1 separate sheet have to do with it? The examples in a leaf do not describe the whole year evenly, but for example 2 examples in January, 27 in February and 555 in December.
If we take the balance line from all leaves as a basis, then in December for this leaf will obviously be the main growth and the deviation from a straight line will be very strong.
If we take the balance line from only this 1 leaf as a basis, uniformity can be achieved, but the participation of overall uniformity is difficult to determine.
Of course we are dealing with intervals, and the smaller we take it, the greater the chance that there will be extremely few examples. There needs to be some balance of reasonableness on this issue, I decided at that point that a year would be optimal for the sheet to show its effectiveness. It is generally normal that in some months there will be no signals at all, especially if there are predictors describing the upper TFs.
Combining leaves into ensembles is a separate task.folk wisdom says you can't see the forest for the trees. But I wonder if you can see the tree by looking through the leaves? I'm not asking about the forest.
Is this the only algorithm you know? Or is it the most efficient? Why do you fixate on it?
It's a passing thought.
Good luck.
folk wisdom says you can't see the forest for the trees. I wonder if you can see a tree by picking leaves. I'm not asking about the forest.
Is this the only algorithm you know? Or is it the most efficient? Why are you fixated on it?
It's a passing thought.
Good luck
1) rules can be extracted from the wooden ones and statistics of each can be calculated, from the HCs not
2) Wooden people are quick learners, NS are not.
Who would have thought that when you know the context, you can trade even on moving averages)))))
Entry and exit prices are calculated by mashka and ohlc and nothing else, who would have thought? I certainly did not... but everything comes with experience.
The brain is the strongest MO (so far), remember that.
Who would have thought that when you know the context, you can trade even on moving averages)))
the entry and exit price is calculated by mashka and ohlc and nothing else, who would have thought? I certainly did not. but everything comes with experience.
The brain is the strongest MO (so far), remember that.
From extreme to extreme?...
Totally unseemly. At least it's down a couple of points.