Forcasting the price, output reading problem
RaginBrain opened this issue · comments
Im am trying to make simple Bitcoin price forcasting NN.
Time interval is 5 minutes.
network has 4 inputs -- 6hidden layer - 1 output it is taking the following data:
[ Price, Volume, Average_price_of_last_5_intervals, Average_volume_of_last_5_intervals]
All inputs are floats and they are all normalized (between 0-1)
output is [Price of next interval] (which changes each iteration)
I have data for training( cca 500 data rows)
Code iterates for each training data row and does the following:
myNetwork.activate([P_1, Vol_1, Vol_5, P_5])
myNetwork.propagate(learningRate, [desired_result]);
After that I try to get some testing results but I cannot read the value of output layer:
console.log(myNetwork.activate([P_1, Vol_1, Vol_5, P_5]))
Am I missing something , is there another way to read output layer?
z = net.activate([ inputs ...
log(Number(z[0]));
@RaginBrain Did you success with your prediction? Can you share a demo?