Bayesian regression - Has anyone made an EA using this algorithm? - page 55
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
I don't get it. To me, you can't compare it to the R at all.
I'm not comparing. Definitely a different environment. Intended mainly for mathematical modelling. In my opinion, it might be quite interesting too, including in our field. Say, for strategy development, modeling and testing. I used to do it in MathLab, but how much does MathLab take up on disk? But here we have packages, structured like in R, with great functionality.
I haven't figured it out yet, but Scilab seems to be able to interact with R. I'm considering Scilab from this very perspective, in cooperation with R. Hypothetically so far. Attracted by the possibility of modeling time series and their processing.
I'm not comparing. Certainly a different medium. Intended mainly for mat modelling. In my opinion, it can also be very interesting, including in our field. Say, for strategy development, modeling and testing. I used to do it in MathLab, but how much does MathLab take up on disk? But here we have packages, structured like in R, with great functionality.
I haven't figured it out yet, but Scilab seems to be able to interact with R. I'm considering Scilab from this very perspective, in cooperation with R. Hypothetically so far. Attracted by the possibility of modeling time series and their processing.
Tip.
Spit on everything, including Scilab along with matlab, and get into R. It's a very deceptive system. At first stages it's very simple, and then you find out there's everything and a bit more. Including matlab.
There are comparisons of R and other similar systems with which it makes sense to compare it. It comes in at the top three. There is a paid version of R, which was acquired by Microsoft. It has started to have add-ons for specific industrial applications, such as processing very large arrays. R is now a standard in statistics, machine learning. Time series is a small part of it. Modern statistical publications almost always include R text besides formulas. Plus it has a very powerful graphics system - it's a piece of cake to draw something, including cartooning extravagances. That is the rule of good mauvais ton. Plus there's an enormous amount of literature. Here, for example, for very limited money.
Tip.
Spit on everything, including Scilab along with matlab, and get into R. It's a very deceptive system. At first stages it's very simple, and then you find out there's everything and a bit more. Including matlab.
There are comparisons of R and other similar systems with which it makes sense to compare it. It ranks in the top three. There is a paid version of R, which was acquired by Microsoft. It has started to have add-ons for specific industrial applications, such as processing very large arrays. R is now a standard in statistics, machine learning. Time series is a small part of it. Modern statistical publications almost always include R text besides formulas. Plus it has a very powerful graphics system - it's a piece of cake to draw something, including cartooning extravagances. That is the rule of good mauvais ton. Plus there's an enormous amount of literature. Here, for example, for very little money.
My dear Sanych, I respect you and I always read your comments carefully, but ... I think the most important thing in the automated forex trading is creating a working model and the rest is technique. I've read several books about R, but I don't understand how it can help me to build a model of a trading system. R works great in data processing, searching for correlations in genetics, biology, sociology, in advertising, political science, etc. But the main thing is that in all these applications of R, there already exists a model, it only needs to be confirmed, clarified, refuted, highlighted.
If I created a working model of forex trading system, then I can calculate the simplest statistical estimations of dispersion, Pearson or chi-square type. Why do I need it? But I still don't see how to build a model with R. Maybe I'm looking in the wrong place.
If I have created a working model of forex trading system, then I can calculate the simplest statistical estimates such as variance, Pearson, or chi-square. The question is why do I need it?
I understand that you can do it all yourself. I do not understand, why do I have to do it myself? Because it all already exists and has already been done. The principle of modern programming - the maximum possible reuse of code, ie what has already been done by others, but not the reinvention of all the same bikes.
sibirqk:
But I still don't see how to build a model with R. Maybe I'm looking in the wrong place.
I don't get it either. MathLab - I understand it, Skilab - poorly, but I also understand how. But I don't understand R.
But ... I think the most important thing in the automated forex trading is to build a working model, the rest is just a technique. I've read several books about R, but I don't understand how it can help me to build a model of a trading system. R works great in data processing, searching for correlations in genetics, biology, sociology, in advertising, political science, etc. But the most important thing is that in all these applications of R, there already exists a model, it only needs to be confirmed, clarified, refuted, highlighted.
If I created a working model of forex trading system, then I can calculate the simplest statistical estimations of dispersion, Pearson or chi-square type. Why do I need it? But I still don't see how to build a model with R. Maybe I'm looking in the wrong place.
R itself is divided into two parts: procedural (algorithmic) programming language and about 8 000 packages with 120 000 functions, which extend the functionality of R language itself. Although I like R language better than many other languages, I think it's senseless to discuss and compare its merits, because the main question: "how much profit will be increased by switching from the metaquote languages to R" remains unanswered.
As for the packages...
When you download R, a number of "basic" packages are installed along with it, which contain statistics and graphics not included in the language itself. Here is what you named.
But the main advantage of R is in the other packages.
Here there is a grouping of packages on different topics, of which there are many more than you mentioned.
I would like to draw your attention to three groups:
There are packages in these groups that are extremely useful in making models for trading. Depending on WHAT you trade, models can be divided into predictive value (regression models) or predictive direction (classification models).
Virtually all trading needs are covered in the approximately 180 packages covered in the caret shell. There you will find both regression and classification as well as tools for preparing raw data (data mining) and evaluation of simulation results, which are much wider than the tester.
I recommend starting with the GUI rattle. Its use is described in my article. The article can be used as a tutorial and there is also an excessively large file attached that you can practice on.
I advertise rattle all the time. It is a very useful system for beginners as it allows you to have the results of 6 models within almost an hour and also covers the entire modelling cycle: data mining-model-evaluation. Besides all your actions are recorded in R-log that may be later used for training and practical work.
Also rattle is very useful for more skilled people, as it allows you to verify ideas very quickly and without errors. When you consider that the main problem in modelling is not the model itself, but the selection of input data (predictors), then rattle becomes extremely useful.
Take up R. It will give professional training in trading for life
Good luck.