Discussion of article "Gradient boosting in transductive and active machine learning" - page 3

New comment
 

Hi Maxim,

Is the learning of the model done only one time at the time of training or is the learning of model going on during live trading as well?

I mean is the model learning itself while it is placing live trades if it places any loosing trades? Is it "Active machine learning" or Am I wrong in understanding?

Thanks

 

Hi Maxim,

Thank you for the English version. I have 3 questions regarding specific parts of the code and I will appreciate if you can answer the questions specifically which will be helpful since I am a basic level programmer and still finding it difficult to understand everything from the explanation.

1.May I know from where and how did you get the below numbers and are these applicable for only "EURUSD" pairs or all currency pairs?

double catboost_model(const double &features[]) { 

    unsigned int TreeDepth[161] = {6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, };

    unsigned int TreeSplits[966] = {393, 730, 93, 54, 352, 313, 540, 591, 217, 12, 576, 757, 208, 574, 756, 446, 505, 10, 487, 791, 210, 673, 125, 647, 286, 593, 523, 706, 566, 510, 575, 754, 325, 450, 470, 321, 438, 589, 48, 257, 283, 745, 707, 520, 564, 296, 702, 27, 524, 223, 404, 755, 60, 218, 387,  };

    unsigned int BorderCounts[20] = {36, 44, 40, 41, 42, 40, 30, 30, 36, 35, 43, 45, 27, 37, 52, 55, 45, 40, 43, 38};

    float Borders[799] = {-0.0103283636f, -0.00538144633f, -0.00438116584f, -0.00384822348f, -0.00290416228f, -0.00226776977f, -0.00186691666f, -0.00173427281f, -0.00136242132f, , -0.00866030902f, -0.0083276052f, -0.00821269862f, -0.00758890808f, -0.0072928248f, -0.00716711534f, -0.00640411209f, -0.00561416801f, -0.0053433096f,  };

2.May I know from where and how did you get the below numbers and are these applicable for only "EURUSD" pairs or all currency pairs?

 

/* Aggregated array of leaf values for trees. Each tree is represented by a separate line: */

    double LeafValues[10304] = {

        -0.02908022041210655, 0, -0.005608946748068618, 0.005129329514937164, 0.03600027378169195, 0, 0.02578289590577986, 0.09444611655822675, 0.03646431117733154, 0.09977346533319338, -0.05595880296318598, -0.069314407568676, 0.08718389822649918, -0.1200338438496052, 0.0693147185156002, 0.01000834600443637, 0, 0.06059264820464737, ,

3.Can you please precisely tell me which parts of the code I need to edit to make it work  for other currency pairs or what exactly I need to do to test it for other pairs?


I have tried with other pairs , but I am not sure if I am doing something wrong or results are simply bad for other pairs where as it is working fine for EURUSD pair. I will appreciate if you can just post another example of some other currency pair to get a better idea how and what to implement to make it work for other pairs.

 
Awesome, thank you! of course i DID take the chance to use your method of model export to mql... great results on new data!
123
New comment