CS 451 Quiz 11
Advice for Applying Machine Learning, Bias vs. Variance
Email address *
Which of the following is NOT a reasonable thing to try if a learning algorithm performs poorly on the TRAINING set? *
Learning algorithms are often trained and evaluated using just two sets (training and test). In what situation is it important to use three sets (training, cross validation, and test)? *
The reason we use both cross validation and test sets is to allow us to tune a subset of parameters on each *
If we split the data into training, cross validation, and test sets, which of the three is usually the largest? *
Using the cross validation set in order to find the best degree of a polynomial to fit the training data is an example of *
If we choose the degree of a polynomial using the cross validation set, the cross validation error will likely be *
Which of the following are indications that an algorithm suffers from high bias? Check all that apply *
Required
High bias means overfitting, high variance means underfitting *
Which data set do we use to tune lambda to get a good balance between underfitting and overfitting? *
Which of the following is NOT a reasonable thing to try if a learning algorithm performs poorly on the TEST set? *
Submit
This content is neither created nor endorsed by Google. Report Abuse - Terms of Service - Additional Terms