Machine Learning Quiz 1

Not a special type of feature normalization

feature scaling

If feature scaling changes the scale of the variables, what does normalization change?

Distribution

One-hot encoding is used for?

categorical variable

If a model requires normalized data, what variables should be normalized?

continuous variables

If a machine learning model requires features to be in a fixed ranges, which scaling method should you use?

Min-max scaling

This regression technique performs both variable selection and regularization is

Lasso regression

This regression technique corrects overfitting and multicollinearity in machine learning models

Ridge regression

Which approach finds regression coefficients in one shot

Least squares

In multiple regression with 3 predictors, the number of beta coefficients is

4

Regularization methods are mainly used to avoid

Overfitting

Parametric methods involve

assuming a functional form for f, then fitting the model

R-squared measures

The proportion of variability in Y explained by X

Mean Squared Error (MSE)

Average squared difference between observed and predicted values

The correct decomposition of the test MSE

Bias² + Variance + Irreducible error

When comparing models with both training and testing data, you should pick the one with the lowest

Test MSE

Overfitting occurs when

the training error is low but test error is high

Are features the same as raw data collected?

No

Benefits of feature selection

Improved performance and interpretability

Which feature selection method evaluates predictor relevance independent of a machine learning algorithm

filter

forward selection is an example of which feature selection method category

wrapper

embedded feature selection methods occur?

during model fitting

when someone uses machine learning to understand how predictors affect the response, they are using it for

interference

Using machine learning solely to forecast future responses is an example of 

prediction

a computational procedure or set of rules used to learn parameters from data is called

An algorithim

Names that refer to inputs include

Independent variables