Here are basic of aaneuromacd:
Inputs:
1. iRSI(NULL,0,14,PRICE_CLOSE,i)/100;
2. iMACD(NULL,0,ma1,ma2,ma3,PRICE_CLOSE,MODE_MAIN,i);
3. iMACD(NULL,0,ma1,ma2,ma3,PRICE_CLOSE,MODE_MAIN,i+1);
4. iMACD(NULL,0,ma1,ma2,ma3,PRICE_CLOSE,MODE_MAIN,i+2);
5. iMACD(NULL,0,ma1,ma2,ma3,PRICE_CLOSE,MODE_MAIN,i+3);
6. iMACD(NULL,0,ma1,ma2,ma3,PRICE_CLOSE,MODE_MAIN,i+4);
7. iMACD(NULL,0,ma1,ma2,ma3,PRICE_CLOSE,MODE_SIGNAL,i);
8. iMACD(NULL,0,ma1,ma2,ma3,PRICE_CLOSE,MODE_SIGNAL,i+1);
9. iMACD(NULL,0,ma1,ma2,ma3,PRICE_CLOSE,MODE_SIGNAL,i+2);
10. iMACD(NULL,0,ma1,ma2,ma3,PRICE_CLOSE,MODE_SIGNAL,i+3);
11. iMACD(NULL,0,ma1,ma2,ma3,PRICE_CLOSE,MODE_SIGNAL,i+4);
outputs:
1. iMACD(NULL,0,ma1,ma2,ma3,PRICE_CLOSE,MODE_MAIN,i) - iMACD(NULL,0,ma1,ma2,ma3,PRICE_CLOSE,MODE_MAIN,i-5)
2. iMACD(NULL,0,ma1,ma2,ma3,PRICE_CLOSE,MODE_SIGNAL,i)-
iMACD(NULL,0,ma1,ma2,ma3,PRICE_CLOSE,MODE_SIGNAL,i-5)
where ma1/2/3 can be set at any value you want
Target prediction is future 5 bar momentum MACD value..
Neural net configuration : back propagation
after training with NS2 then you can display future 5 bar MACD as predicted value + current MACD.