These questions were used in gadsdc1 but got removed from gadsdc2; they could be reintroduced somewhere, perhaps:
Suppose you have features that correspond to length, width, and height, and you're predicting a label which is volume. Could you use KNN? Explain how it could work (with some relevant implementation details), or why it wouldn't work.
Suppose you have features that correspond to length, width, and height, and you're predicting a label which is volume. Could you use a Bayesian approach? Explain how it could work (with some relevant implementation details), or why it wouldn't work.
Suppose you have features that correspond to length, width, and height, and you're predicting a label which is volume. Could you use linear regression? Explain how it could work (with some relevant implementation details), or why it wouldn't work.
What determines the complexity of a linear model, as it relates to model bias and variance? How can you control this complexity?
These questions were used in gadsdc1 but got removed from gadsdc2; they could be reintroduced somewhere, perhaps: