Machine learning in trading: theory, models, practice and algo-trading - page 1309
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
Stubbornness is close in meaning to perseverance. I hope they will help you get to the point of successfully implementing your ideas in the ME. These are useful qualities for researchers. ;-)
PS I came up with a name for your leaf selection system: "Herbarium" - add to the collection of methods from trees, forests, stumps, jungles.Being stubborn is not always productive, but thanks for appreciating my quality.
As for the herbarium, I already use that terminology, thank you for supporting the idea
Forum on trading, automated trading systems and testing trading strategies
Gathering a team for the development of MO (Decision Tree/Forest) for trend strategies
Aleksey Vyazmikin, 2018.06.27 16:42
A little bit about the directions, which I see promising:
1. Work on genetic methods - here I want to try to create logical groups of features, including when sowing and nurturing offspring.
2. Work on fitness functions - here I want to provide feedback on the actual results of the trade, not just the results of classification.
3. Work on stumpage nurturing - a semi-automatic method of generating trees when the base of the tree is known or when individual branches are known.
4. Work on a method for evaluating the resulting tree.
5. Working on a convenient and fast way to integrate the tree into the EA's algorithm.
6. Work on creating a herbarium. - By pulling individual rules to create scaffolds.
7. Work on a method of building a tree that does not exclude limits/ranges, but tries to give solutions to all values in the predictor (relevant when quantization (partitioning into subgroups) is already used in preparing the predictor).
I've figured out a little with drawing graphs, now I can quickly look at the models, pre-selected by the script by condition, visually, for now I decided to save only graphs outside of training.
This is a gif, but it only shows frames when I click on it - I don't know what's the reason...
...
It's a gif, but I only get frames when I click on it - I don't know what the reason is...
Make it a smaller size. The width should not be more than 750 pixels.
Make it smaller. No more than 750 pixels.
I see, I will take it into account next time, thank you.
I've figured out a little with drawing graphs, now I can quickly look at the models, pre-selected by the script by condition, visually, for now I decided to save only graphs outside of training.
This is a gif, but I have it showing frames only when I click on it - I don't know what's the reason...
If you have a year of OOS, then purely psychologically it will be difficult to sit in a drawdown or in a chattering about 0. And if there are investors, for this period many may run away.
Graph 3 - it seems without such unpleasantness.
If you have OOS for a year, then purely psychologically it will be difficult to sit in a drawdown or in a chatter around 0 for about half a year.
Well, there are different options on the gif and even they are not to demonstrate a successful model, but just about the development of a project to automate the learning cycle. Yes, it's almost a year, but if you think that I could drain that money to zero, it's better to have a small profit and keep working. Besides, the ideology of my using different models at the same time - for each model there will be a fixed volume - I'll work in two accounts at once - buy and sell.
I will work in two accounts at once - buy and sell. Now (soon, I hope) I will finish finish predictors and try some other strategies with them, but so far the list is not so big.
And this is what the balance of the classification error of the same models looks like, i.e. if correctly classified 1, then +1, if classified 1, and there was zero, then -1.
And this is the balance of classification error of the same models, that is, if correctly classified 1, then +1, if classified 1, and there was zero, then -1.
it's time to organize a convention of MOs, after everyone will have a first million purely from the MOs (minimum)
I think, in a year, some will have
let's rent a conference hall, god forbid, hennessy, then a yacht tour
It's time to organize a convention of MO guys, after everyone will have the first million purely from the MO profit (minimum)
I think in a year some of them will have it.
We'll rent a conference room, sorry... God, a Hennessy, then a yacht tour.
My aims are very modest - ideally would earn thousand 200 rubles this year, to not die of hunger...
My goals are very modest - I would ideally like to earn about 200 thousand rubles during this year in order not to starve...
I understand that an artist should be hungry, but it's better to do it in your spare time)
As an option, if there are no problems with going to work, there are many places now that require datascientists from the PZ good