MDFT Academy Quizzes

How can you fix this problem?

You are building a regression model to predict the wind speed at any point on earth during any time of the day. You are training your model with Gradient Descent and notice that the loss value does not decrease smoothly during training. Instead, it changes chaotically and the training often terminates with a low-qualty model.

How can you fix this problem?

Choose the correct answer from the options below.

I'm sorry, your answer is not correct.

Increase the 'number of epochs' hyperparameter is incorrect. Number of epochs specifies how long the model is trained. Increasing the training time will not solve the problem.
Decrease the 'number of epochs' hyperparameter is incorrect. Number of epochs specifies how long the model is trained. Descreasing the training time will not solve the problem.
Increase the 'step size' hyperparameter is incorrect. Increasing the step size will not fix the problem. Gradient descent is wildly jumping around on the loss surface because the step size is too large.
Decrease the 'step size' hyperparameter is correct

2. Regression

How can you fix this problem?