Question: 2. (40 marks) Using Feedforward neural networks (FFNN): (a) Train a two-hidden-layer FFNN using the training data. You are free to de- termine or just

2. (40 marks) Using Feedforward neural networks
2. (40 marks) Using Feedforward neural networks (FFNN): (a) Train a two-hidden-layer FFNN using the training data. You are free to de- termine or just use the default values of the (hyper-) parameters. You are sug- gested to try different settings to see the impacts. However, one implementation suffices for this question while you can simply use the settings in the table below. (b) Use the trained model in part (a) to predict for the test data (X.test). Note that the predicted values are transformed returns. Hence, you need to first convert it back to returns (R.pred ) by reversing the formula in Q1(b) and then use the return to recover the price prediction: S.pred, = St-1 * (1 + R.pred;). (c) Compute the test mean absolute percentage error (MAPE) by the formula 1 St - S.predt n.test St where n.test is the number of test data and the summation loops through the dates for the test period. number of neurons in hidden layers 128, 128 activation functions ReLU or Sigmoid (for FFNN), Tanh (for LSTM) optimizer by default or RMSprop batch size 16 the number of epochs 30

Step by Step Solution

There are 3 Steps involved in it

1 Expert Approved Answer
Step: 1 Unlock blur-text-image
Question Has Been Solved by an Expert!

Get step-by-step solutions from verified subject matter experts

Step: 2 Unlock
Step: 3 Unlock

Students Have Also Explored These Related Accounting Questions!