Machine learning in trading: theory, models, practice and algo-trading - page 569
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
They recently wrote that they are going to do it, I don't remember which topic.
If you want to trade on the exchange, you may trade in the kitchen... And the tester is cool... Python or a couple of libs with machine learning and you don't need anything else to live on.
A year ago I first applied to enthusiasts, and then there was a month order in the Market to modify for MT5 a ready-made library for R. There was exactly one performer who knew what to do and a lot of "professionals" who demonstrated their prowess and obvious desire to learn for my money.
Where were you?
Maybe you will show some altruism and make a bundle with R on the server-client scheme?
Or maybe (can you imagine!) make a real link, where the R core is like a dll for µl? This variant is a notable development in the world, with all the consequences for the implementer.
I will not, SanSanych. I do not work to order, only for myself. And in general I do not like to program - only on demand. And so all my life.) All my life I don't like it, and all my life I've been programming.)
But if I happen to need R, I promise I will not hide it.
I don't have any problems with MT terminal, believe me - I've been trading since 2008. For Forex it is the best.
I don't know, there were no problems at the Moscow one
I don't know, I didn't have any problems at the Moscow one.
Try Quick and feel the difference. I trade a lot of hands. I also trade options. Once again, Quick is not the best and rather outdated terminal.
By the way, I'd say that Lua-C++ with callback functions provides a lot of useful information.
Try Quick and you'll feel the difference. I trade a lot of hands. I also trade options. Once again, Quick is not the best and outdated terminal.
By the way, I've tried it and it pisses me off.
When I see something ugly and outdated, I can't work with it :)
when i was studying i had 1C accounting... that's the end of the line
I used to have 1C in my class, so I had no idea what to do with it.)
I've tried it, it pisses me off )) when I see something ugly and outdated, I can't work with it :)
When I was studying, I had 1C accounting... that's the end of the line
When I was studying, I had 1C in accounting and that was the end of it :) thinkorswim is nice, but I have only one place to trade with.
At first, yes, it was annoying. But it's no better now. I have never been used to trade with Quickwim.
I had never been used to trade on it, I had never tried it before.) It's a pity my old terminal was killed - it was the best. It was the best one. The developers have gone, but new ones couldn't cope with updates.
About my library. It has a problem. Establishing a working directory. Google doesn't help. The methods of Windows installs a working directory in the library, but Python does not pick it up. In general, the library needs improvement.
For cloud enthusiasts
Services and tools for building intelligent R applications in the cloud
For cloud enthusiasts
Services and tools for building intelligent R applications in the cloud
About Azure - it looks quite tempting at the 1st stage, but there is a very limited number of free calls to the built model through API, and with a paid subscription it will be economically viable only for commercial use, i.e. for large projects. I'm not sure about other services, but I think it's about the same.
Just for experiments Azure is pretty handy, there are a number of Built-in models, including preprocessing, model comparison, all in the form of block diagrams. And the rest can be written in R or Python and connected as modules via block diagrams, and for free.
https://medium.com/@ceshine/feature-importance-measures-for-tree-models-part-i-47f187c1a2c3
I should try to add at least 1 method to the algib forests and then everything can be done automatically in MT5 without R, for example, data retrieval
i can do it automatically without R. e.g., data retrieval. somewhere else they write the easiest way - shuffle values in each attribute and retrain again if the quality drops much, then the variable is significant... but i think this is not the right way, and the sources are questionable.
SanSanych, how do you know which method is used to calculate R?
i'm sure that in the forex market the importance will "float" very strongly from set to set, but still fun