I am new to LSTM and was trying to do a many-to-one multivariate prediction where I have 6 inputs values and my 7 th input are my targeted values. My model is as follows:
# create and fit the LSTM network
model = Sequential()
model.add(LSTM(128,activation='relu', input_shape=(trainX.shape[1],trainX.shape[2]),return_sequences=False))
model.add(Dropout(0.2))
model.add(Dense(1))
model.compile(loss='mse', optimizer='adam', metrics=["mse"])
model.summary
When I plot the history of the model, I find that the validation values are so small (i.e loss of less than 0.005) and the validation history fluctuates. What does this mean and what advice would you provide for a more efficient multivariate prediction with a single output?