Machine Learning in MT4 & MT5

New comment
 

Hello Dear MetaTrader users,


Please do let me know if anyone can give me a brief idea on how to apply Machine Learning concepts in MT4 & MT5 using MQL Language.


Thanks

 
Alex Grischuk:

Hello Dear MetaTrader users,


Please do let me know if anyone can give me a brief idea on how to apply Machine Learning concepts in MT4 & MT5 using MQL Language.


Thanks

Well, at least this is how i did. //--->

#1 --> I created a "BUY/SELL" button on chart using EA.

#2 --> Whenever I opened the trade using the button created earlier, it will grab and save a few parameter (ie. Close,Open,High,Low) into a text file.

#3 --> Then I created an apps using Matlab/Python (i used both) to grab this ".txt file".

#4 --> The apps will do an analysis based on the data in ".txt file" and at the same time build a neural structure.

#5 --> Instead of that, the apps also will write the predict output from the current neural structure into another ".txt file".

#6 --> Then, the EA from earlier will recognize this new ".txt file" and display the output on the chart.


//It's easy isn't it? You just need to have a correct parameter, or else your neural structure will be useless.

//More trade I made, more data I have. Which mean more complex the neural structure will be, thus leading to more accuracy.

Goodluck~

FileWrite - File Functions - MQL4 Reference
FileWrite - File Functions - MQL4 Reference
  • docs.mql4.com
The function is intended for writing of data into a CSV file, delimiter being inserted automatically unless it is equal to 0. After writing into the file, the line end character "\r\n" will be added. Numbers will be converted into a text at output (see the Print() function). Data of the double type are output with the accuracy of 16 digits...
 
Alex Grischuk:

Hello Dear MetaTrader users,


Please do let me know if anyone can give me a brief idea on how to apply Machine Learning concepts in MT4 & MT5 using MQL Language.


Thanks

Look at this search results links (for the articles).
Search - MQL5.community
Search - MQL5.community
  • www.mql5.com
Searching is based on morphology and is insensitive to case. All letters, no matter of their case, will be processed as lowercase. By default, our search engine shows pages, that...
 
Sergey Golubev:
Look at this search results links (for the articles).

Thanks Sergey

 
Ahmad Zuhairdi Noh:

Well, at least this is how i did. //--->

#1 --> I created a "BUY/SELL" button on chart using EA.

#2 --> Whenever I opened the trade using the button created earlier, it will grab and save a few parameter (ie. Close,Open,High,Low) into a text file.

#3 --> Then I created an apps using Matlab/Python (i used both) to grab this ".txt file".

#4 --> The apps will do an analysis based on the data in ".txt file" and at the same time build a neural structure.

#5 --> Instead of that, the apps also will write the predict output from the current neural structure into another ".txt file".

#6 --> Then, the EA from earlier will recognize this new ".txt file" and display the output on the chart.


//It's easy isn't it? You just need to have a correct parameter, or else your neural structure will be useless.

//More trade I made, more data I have. Which mean more complex the neural structure will be, thus leading to more accuracy.

Goodluck~

Can you show more details? I would like an example to be used in strategy Tester.  Thanks 

New comment
Reason: