Hi,
I'm trying to use a custom Objective for my advise functions, but right now my adviseLong always returns zero.
My Zorro code looks like this
var params[10];
params[0] = cubicDev10N[0];
params[1] = deltaPVR5N[0];
params[2] = deltabWidth3N[0];
params[3] = MMIFasterN[0];
params[4] = deltaMMIFastest10N[0];
params[5] = HurstFast[0];
params[6] = HurstFaster[0];
params[7] = ibsOneN[0];
params[8] = ATRSlowN[0];
params[9] = marketVolNorm[0];
set(RULES);
var myObje = ifelse(priceClose(-1) - priceClose(0) > 20,1,-1);
LifeTime = 1;
MaxLong = 1;
if(Train)
{
Hedge = 2;
set(PEEK);
adviseLong(NEURAL+BALANCED, myObje, params,10);
}
if(!Train)
{
var LongPrediction = adviseLong(NEURAL, myObje, params,10);
printf("nPred for long is %f", LongPrediction);
if(LongPrediction > 0.5)
{
enterLong();
}
}
If I change my code to use SIGNALS during training, and give the resulting input to my neural.predict function straight from R, the predictions are returned as expected. So I must be doing something wrong on the Zorro side.
I've trying putting enterLong() in the if(Train) block as well, but that doesn't make any difference, and I don’t think it's needed when not using trade results as the target?
Anything obviously wrong with Zorro my code?
