Machine learning in trading: theory, models, practice and algo-trading - page 3065

 
fxsaber #:

Scalpers - interesting.

Scalpers are expensive to do on regular hardware for now, but I'll give it a try. Model training is a bottleneck, slow on big data, and then the response is relatively long.
 
mytarmailS #:

Vladimir, did you use the blotter package?

by the way, the strategy from this book really works on the day trades, on stocks but not on forex.


and so on. A lot of stocks...


Although I used data from a brokerage centre that has stocks, so everything is possible ...

7.7 Evaluating Trading Rules | Techincal Analysis with R
  • Ko Chiu Yu
  • bookdown.org
This is an introductory textbook that focuses on how to use R to do technical analysis.
 
Maxim Dmitrievsky #:
derivatives are traded on the exchange, bitcoin doesn't care about that.

I was talking about the cryptocurrency exchange of course.....

 
Aleksey Vyazmikin #:

I was talking about the cryptocurrency exchange of course.....

Without it, you can't exchange the tokens
 
Maxim Dmitrievsky #:
Without it, you can't exchange signs

There are exchanges. But, in general, something about something else already... I mean that tokens have an original meaning, which is lost over time and only the exchange rate remains. And since there's no economic sense in them, their price is just speculative. It's clear that bitcoin is now being used essentially for dark deeds - that's its value.

 
Aleksey Vyazmikin #:

What is the sr - why isn't it on the list that was given? What are you using it for?

The sr package is "
... Selecting good prognostic inputs using gamma test ...

... to determine lags in the time series. How far ahead does the signal appear? This is a classic example of the problem of figuring out which of our input variables, the predictor variables, is actually predictive. This problem is nonlinear and multivariate. This is the problem that the Gamma test solves.

...The use of Gamma to find causal relationships in time series data, determine optimal delays and attachments, and tune the performance of a neural network is demonstrated. "

Missing from the list. Purpose of use above.

 
mytarmailS #:

hopelessness )))

Absolutely.

The blotter package did not apply. It is not in CRAN at the moment.

 

Asked Kandinsky to draw a trader with packages


 
Vladimir Perervenko #:

Package sr - "
... Selecting good predictive inputs using gamma test ...

... identifying lags in the time series. How far ahead does the signal appear? This is a classic example of the problem of figuring out which of our input variables, the predictor variables, is actually predictive. This problem is nonlinear and multivariate. This is the problem that the gamma test solves.

...The use of Gamma to find causal relationships in time series data, determine optimal delays and attachments, and tune the performance of a neural network is demonstrated. "

Missing from the list. Purpose of use above.

Should be a useful thing, judging by the description.

How to run this for testing on a sample - can you share the code for the public?

 
Aleksey Vyazmikin #:

Should be a useful thing, judging by the description.

How to run it for testing on a sample - can you share the code for the public?

similar to the Kolmogorov-Smirnov test

To test something, you have to first determine what to test. That usually takes forever in our field.

It's desperate.