Discussion of article "Deep neural network with Stacked RBM. Self-training, self-control" - page 2

New comment
 

Dear Vladimir, congrats for your article, very interesting.

I have a question. At the beginning of your article you do:

 

> data.f <- data[ ,-highCor]
> colnames(data.f)
 [1] "DX"     "ADX"    "oscDX"  "ar"     "tr"    
 [6] "atr"    "cci"    "chv"    "sign"   "vsig"  
[11] "slowD"  "oscK"   "signal" "vol"    "Class"

which shows the column "Class" and then you use it in prepareBest:

 

idx <- rminer::holdout(y = data.f$Class)

 

But there is no evidence of this column "Class" in the "In" function.

From where does it come from?

Thanks. Cheers. 

 
mg64ve:

Уважаемый Владимир, поздравляю для вашей статьи, очень интересно.

У меня есть вопрос. В начале вашей статьи вы делаете:

который показывает столбец "класс" , а затем использовать его в prepareBest:

Но нет никаких доказательств этой колонке "Класс" в функции "In".

Откуда она взялась?

Благодарю. Приветствия.

\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\

Hi, 

Смотрите функцию form.data(). В ней объединяются предикторы x <- In()  и целевая y <- ZZ() %>% as.factor(). "y" расположен в последнем столбце  data.frame и является целевой переменной. Я не помню где она переименовалась в переменную "Class", но это та же целевая "у".

 

Hi, 

See form.data () function. It combined predictors x <- In () and target y <- ZZ ()%>% as.factor (). "Y" is located in the last column is data.frame and is target variable. I do not remember where it was renamed in the variable "Class", but it is the same target "y".

 

Hi,

 Thanks for the great sharing. Which factors could cause  Alert("Rterm crash!(OnTimer)")? Since Rterm is successfully established in init() phase with correct hR returned. I have download the R 3.2.2 version but it won't be called up when I run the EA. It shouldn't be one of the results after acquired by Mircrosoft, I hope.  

 

After downloading the RStudio as suggested, there are still some errors, like "Error: object 'first' not found"  when to run "e_DNRBM.r". I suppose this should be an entry question but still want your suggestions.  

 

 Thanks for your explanation in advance!

Peter 

 
Peidong Zhu:

Hi,

 Thanks for the great sharing. Which factors could cause  Alert("Rterm crash!(OnTimer)")? Since Rterm is successfully established in init() phase with correct hR returned. I have download the R 3.2.2 version but it won't be called up when I run the EA. It shouldn't be one of the results after acquired by Mircrosoft, I hope.  

 

After downloading the RStudio as suggested, there are still some errors, like "Error: object 'first' not found"  when to run "e_DNRBM.r". I suppose this should be an entry question but still want your suggestions.  

 

 Thanks for your explanation in advance!

Peter 

Hi Peter,

The main reasons for the fall Rterm:

- Lack of the necessary libraries

- Loaded libraries of more recent versions are not compatible with the script.

- Incorrect path to the scripts.

To determine the cause:

- Rstudio load the attached file  e_Sess_DNRBM_GBPUSD_30.RDATA (Sess_DNRBM_GBPUSD_30.zip).

- Load the library manually.

- Run the script  e_DNRBM.r  line by line with a line 23.

- Not save.image(file = fS)

To determine where the error is and what. 

Best regards

Vladimir 

 
Vladimir Perervenko:

Hi Peter,

The main reasons for the fall Rterm:

- Lack of the necessary libraries

- Loaded libraries of more recent versions are not compatible with the script.

- Incorrect path to the scripts.

To determine the cause:

- Rstudio load the attached file  e_Sess_DNRBM_GBPUSD_30.RDATA (Sess_DNRBM_GBPUSD_30.zip).

- Load the library manually.

- Run the script  e_DNRBM.r  line by line with a line 23.

- Not save.image(file = fS)

To determine where the error is and what. 

Best regards

Vladimir

 

 

Hello Vladimir,

 

Thanks for the quick feedback. 

 When I follow your guides to run the script  e_DNRBM.r  line by line, the first error encountered is:   Error in rev(o) : couldn't find the object 'Open'.

How could I handle this problem? It seems the Open price data hasn't been loaded or positioned correctly.

Thanks in advance.

  

Best regards

Peter 

 
Peidong Zhu:

Hi Peter,

This error appears when the performance of a line?

I wrote - to execute the script starting at line 23. 

Best regards

Vladimir 

 
Deer all! After backtest and trading live, i see the H1 time frame is a good selecting for trading. Avoiding flat market is needed, and selecting EURUSD is large risk. Thank Vladimir!
 

Hello Vladimir,

I'm able to run the EA in backtest but it seems to perform only the first pass.

In the attached file you can find the output of both tester and debugview.

Can you please help?

Thanks,

Diego 

Files:
TRANX-CALC.LOG  23 kb
 
lord_hiro:

Hello Vladimir,

I'm able to run the EA in backtest but it seems to perform only the first pass.

In the attached file you can find the output of both tester and debugview.

Can you please help?

Thanks,

Diego 

I think the problem be due to darch package. Using the darch package V 0.10.0 to correct it.
12345678
New comment
Reason: