Machine learning in trading: theory, models, practice and algo-trading - page 3165
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
Recently I see quite a lot of articles with "homemade" trees. As a rule, they do not write trees from scratch, but use the rpart package from R. So one thing does not interfere with another (self-made packages).
There is no analogue of the package in python, I think. So saving a homemade tree in ONNX will surely be problematic.
do you need ONNX for trees ?
Do you need ONNX for trees ?
For those trees that need to be used in an mql5 Expert Advisor, it would not hurt. I am suspicious of the idea of trading from R.
For those trees that need to be used in an EA on mql5, it would not hurt. I am suspicious of the idea of trading from R.
No, that's not what I meant...
trees are log rules, they are relatively easy to transfer to native mql code, it's not neuron or nuclear svm.
No, that's not what I meant.
wood these are log rules, it's relatively easy to port them to native µl code, it's not like it's neuron or nuclear svm.
Yes, I agree, if we are talking about one tree. If you want to make a forest, it's more problematic, though you can also generate c-code, I guess. But with ONNX somehow everything is more practical, especially if you need to keep track of several EAs, periodically retrain models, change hosting, etc.
And how many of you have implemented the ideas discussed here?
I've done a lot of things. The result is about zero.
More and more seriously I think it's time to end this epic with attempts to find regularities in random data. I will code a couple more ideas and test them, and then we will see.
The feeling that knowledge is diminishing, which is the fault of batch thinking.
There are different kinds of knowledge: knowledge sucked out of one's own finger and knowledge of other people's results, who have bothered to formalise this knowledge in the form of a code, repeatedly tested by other people and practice.
For those trees that need to be used in an EA on mql5, it would not hurt. I am suspicious of the idea of trading from R.
There is no problem. In a normal EA, the block of accepting a trade order is replaced by a call to R. The next bar is sent to R, and Long-Out-Short is sent back.
There are different kinds of knowledge: knowledge sucked out of one's own finger, and knowledge of the results of other people who have bothered to formalise this knowledge in the form of a code, repeatedly tested by other people and practice.
Other people's knowledge multiplies and yours decreases. In general it is normal, you cannot know everything. But it is not normal to preach package thinking among those who have more knowledge than you.
Or are you some honoured master in ML, a professor maybe or a cool manager? Perhaps there is something we don't know.Other people's knowledge increases and yours decreases. In general, this is normal, you can't know everything. But it is not normal to preach package thinking among those who have more knowledge than you.
Or are you some honoured master in ML, a professor maybe or a cool manager? Perhaps there's something we don't know.You're just annoying, guru! About me it is not known, but about you it is known for sure - look in your market!
You're just annoying, guru! I don't know about me, but you do - look in your market!
Do you have anything left besides painful perception of the world around you and holy faith in packages, or the piggy bank is exhausted? )
Every day I read about packages you don't know a thing about.
Or is it Groundhog Day?