predict future value in every one hour using (t+60 minutes) LSTM neural network in python

Question

I have a data csv file including with three inputs and two output with time series. Here data took an every one hour one hour. So I need to predict my next future value at t+60 according to the previous input value and at that time period if having new input value using regression neural network. So I choose LSTM neural network to predict next future value. But I don't know how to give time period to predict my future value. Can anyone suggest me how to solve this problem? Can anyone give me any examples to clear out this problem? Here that prediction value will come as input value (g).

subset of my csv file

here I upload my code;

def create_data(data, look_back=1):
dataX, dataY = [], []
for i in range(len(data) - look_back - 1):
    a = data[i:(i + look_back), :]
    dataX.append(a)
    dataY.append(data[i + look_back, 2])
return numpy.array(dataX), numpy.array(dataY)

data = pd.DataFrame(data,columns=['g','p','c'])
numpy.random.seed(7)
data = data.values
scaler = MinMaxScaler(feature_range=(0, 1))
data = scaler.fit_transform(data)
train_size = int(len(data) * 0.67) 
test_size = len(data) - train_size
train, test = data[0:train_size, :], data[train_size:len(data), :]
# reshape into X=t and Y=t+1
look_back = 3
trainX, trainY = create_data(train, look_back)  
testX, testY = create_data(test, look_back)
# reshape input to be  [samples, time steps, features]
trainX = numpy.reshape(trainX, (trainX.shape[0], look_back, 3))
testX = numpy.reshape(testX, (testX.shape[0],look_back, 3))
model = Sequential()
model.add(LSTM(6, input_shape=(look_back,3)))
model.add(Dense(1))
model.compile(loss='mean_squared_error', optimizer='adam')
history= model.fit(trainX, trainY,validation_split=0.33, nb_epoch=10, 
batch_size=30)
trainPredict = model.predict(trainX)
testPredict = model.predict(testX)
trainPredict_extended = numpy.zeros((len(trainPredict),3))
trainPredict_extended[:,2] = trainPredict[:,0]
trainPredict = scaler.inverse_transform(trainPredict_extended) [:,2]  
print(trainPredict)
testPredict_extended = numpy.zeros((len(testPredict),3))
testPredict_extended[:,2] = testPredict[:,0]
testPredict = scaler.inverse_transform(testPredict_extended)[:,2]   
trainY_extended = numpy.zeros((len(trainY),3))
trainY_extended[:,2]=trainY
trainY=scaler.inverse_transform(trainY_extended)[:,2]
testY_extended = numpy.zeros((len(testY),3))
testY_extended[:,2]=testY
testY=scaler.inverse_transform(testY_extended)[:,2]
trainScore = math.sqrt(mean_squared_error(trainY, trainPredict))
print('Train Score: %.2f RMSE' % (trainScore))
testScore = math.sqrt(mean_squared_error(testY, testPredict))
print('Test Score: %.2f RMSE' % (testScore))
# shift train predictions for plotting
trainPredictPlot = numpy.empty_like(data)
trainPredictPlot[:, :] = numpy.nan
trainPredictPlot[look_back:len(trainPredict)+look_back, 2] = trainPredict

# shift test predictions for plotting
testPredictPlot = numpy.empty_like(data)
testPredictPlot[:, :] = numpy.nan
testPredictPlot[len(trainPredict)+(look_back*2)+1:len(data)-1, 2] = 
testPredict

Answer 1

One option is to :

1. Break data into constant-frequency observations (E.g.: Assume that g,p,c,out are same for all time periods between two observations). With this, you will get samples every N minutes.
2. Training data will then be a set of [Last M observations till T, Observation for T + 60]

This Train and test data can be fed into a network.