Machine learning in trading: theory, models, practice and algo-trading - page 1919
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
A picture like this
gives this graph of the profit
how to make a screenshot, that's how it can be saved, and then just from a blank sheet to load and the model appears?
Yes
That's how do you interpret that kind of stacking?
Well it's a cluster, it looks like objects in multidimensional space merged into such clumps in three dimensions, so that you can look at your 100-500 dimensional data in three dimensions, and somehow estimate their structure
By the way, how to set colors to classes forcibly - I don't understand what is what?
there colors are already set through the variable target, you see that there are three colors, that's your target in three classes
By the way, the picture above with the islands is after my method of transforming the predictors, and this is what the original sample looks like, before the transformations
Two nebulae...
The data is essentially the same, but the result, even visually, is different. So I don't know how to use this tool, except as an estimate of the readiness of the sample to learn....
Well, here's a 3D model, say, which opens in a separate window and there's not even a way to make a screenshot, so it can be saved like this, and then just from a blank sheet to load and the model appears?
You can even recognize new data with it
Well it's a cluster, it looks like objects in multidimensional space combined into such clumps in three-dimensional space, so that you can look at your 100-500 dimensional data in three dimensions, and somehow estimate their structure
Maybe the target should be by the number of these islands? I don't understand the qualitative assessment of this visualization.
there should be a way to specify colors - I see that the back is black, and the others are 1,2,3 - how do you translate that into color?
you can even recognize new data with it
What do you mean, what data? Where does it come from?
Maybe the targets should be by the number of these islands? I do not understand the qualitative assessment of this visualization.
Understand that these clusters are an objective reality, the real structure of your data...
your target is a subjective reality, that's why your class labels have nothing to do with the cluster structure
Look at the example with irises and you'll understand everything.
You have to specify the color - the back side is black, but the rest - 1,2,3 - how do you translate this into color?
Look in the manual in the comments there is everything
What do you mean, what kind of data? What data?
)))
Well, like with the usual Forrest nets, new data appeared, they can be recognized by the saved model
Understand that these clusters are an objective reality, the real structure of your data...
Your target is a subjective reality (self-reality), so your class labels have nothing to do with the structure of the clusters.
So, I was wondering, what if we partition this sample into these islands and already inside them do model training.
Different clusters indicate different factors that influence the result, or rather their different weight, I think.
Only how to make this division ...
Here by the way from another angle all the same full data - you see 4 clusters
So I don't know how to use this tool, except as an assessment of sample readiness for training....
Use as intended - visualization of multidimensional space to evaluate data structure, whether there are clusters or it's a point cloud, etc.