Bayesian regression - Has anyone made an EA using this algorithm? - page 54
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
these functions go in the database.
var, cov, cor I already found it myself. :) But the rest, thank you very much.
Document R: A Language and Environment for Statistical Computing is about 3500 pp. :)
And then there are many packages with good examples of use in manuals and added pdf's
That's where I'm looking, with varying degrees of success. :) In a search, you give him the word and he's up to 200 links.
For the hundredth time, I advertise rattle.
The problem is not just finding a particular function, but finding a functionally complete set of functions. And as was rightly noted above, with limited knowledge finding such a set of functions is quite a problem...
If you take rattle, you use a very primitive interface to use a functionally complete set of R tools: data mining, modelling (6 ideologically different models that can be used for regressions and classifications), model evaluation. And in the meantime you get a log in R, in which you can see the tools used... This will be a very limited reference, all to the point. And then already think where to run and where to save yourself.
PS.
If R stands, running rattle is a matter of minutes. You can prepare yourself an input excel file, you can take from the attachment to my article. In a couple of hours you'll see the result, and in another couple of hours you'll have a lot of very informative thoughts....
For the hundredth time I advertise rattle.
Robert I. Kabakov | R in action
rattle - yes, I'll look it up.
I found an interesting book on the web on R -
Robert I. Kabakov | R in action
rattle - yeah, I'll look for it.
Eh, five years ago this book ...
Yeah, I'll look for it.
What's there to look for? The usual package, installs like all the others.
The documentation is the same as for all the other packages here.
If the developers keep their promise and build the API to R into MT4, the connectivity problems will disappear. Even in the status quo, though, there are no unsolvable problems.
The most effective learning is when solving a practical task, a simple one to begin with. For example, an Expert Advisor that trades in a channel. Construct the channel using three curves: Middle - F - Kolmogorov-Zhurbenko filter (package "kza"); upper = F + 2*sd(High - F); lower = F + 2*sd(Low - F). Well, the rule is standard (for starters) - from boundary to boundary. Further you can complicate it to infinity. The algorithm is simple but effective.
This channel looks as follows on the chart. This is the simplest example. It is not intended for discussions.
Good luck
If the developers keep their promise and build the API to R into MT4, the connectivity problems will disappear. Even in the status quo, though, there are no unsolvable problems.
The most effective learning is when solving a practical task, a simple one to begin with. For example, an Expert Advisor that trades in a channel. Construct the channel using three curves: Middle - F - Kolmogorov-Zhurbenko filter (package "kza"); upper = F + 2*sd(High - F); lower = F + 2*sd(Low - F). Well, the rule is standard (for starters) - from boundary to boundary. Further you can complicate it to infinity. The algorithm is simple but effective.
This channel looks as follows on the chart. This is the simplest example. It is not intended for discussions.
Good luck
I wonder where it was promised that R would be attached to MT4?
In one of the threads discussing the expected improvements in the new build of MT5. Can't find it fast enough.
Actually, the API to R from MT doesn't need one for nothing. Not a problem at all, not even an issue. The issue for MT is quite different - it practically doesn't support event-driven programming. A kind of ersatz.
And second, there is no API to the MT itself.
While searching for R docs I found such a thing as scilab. Very interesting environment. Same libraries, as in R, and many other interesting things, including LabView-like. The access is more or less the same as to R. Yes, it's freely distributable.
Actually, the API to R from MT doesn't need one for nothing. Not a problem at all, not even an issue. The issue for MT is quite different - it practically doesn't support event-driven programming. Some kind of ersatz.
And secondly, there is no API for MT itself.
Searching for R docs I've found another thing - scilab. It is a very interesting environment. Same libraries, as in R, and many other interesting things, including LabView-like. The access is more or less the same as to R. Yes, freely distributable.
I don't get it. I don't think it can be compared to R at all.
Here's
Scilab is a computer maths system which is designed
for calculations such as:
-
solving non-linear equations and systems;
-
solving linear algebra problems;
-
- solving optimization problems;
-
differentiation and integration;
-
experimental data processing tasks (interpolation and
approximation,
-
least squares method);
-
solving ordinary differential equations and systems
How can it be compared to a system which is already top-notch in our area, in the area of model development for trading?