Question 1 (1 point)

Which one best describes about the difference between simple and multiple regression?

Question 1 options:

	In simple regression least square is used to estimate the parameters while multiple regression uses maximum likelihood
	In simple regression there is a single predictor while in multiple regression there are more than one predictors
	In simple regression the parameter estimates are averages while in multiple regression the parameters are summed over all the predictors
	Simple regression has an intercept term while multiple regression does not.

Question 2 (1 point)

What does least square minimize? Select all that apply.

Question 2 options:

	The squared difference between actual and estimated values
	The residual sum square
	The difference between the estimated value of the dependent variable and the largest parameter value
	The loss (cost) function

Question 3 (1 point)

Which of the following can be used as accuracy measures for linear regression? Select all that apply.

Question 3 options:

	R-Squared (R)
	Precision & Recall
	F2-Score
	Residual Standard Error (RSE)

Question 4 (1 point)

What does a linear regression model?

Question 4 options:

	The aggregated weights of the independent variables
	The category of the dependent variable
	The log of the dependent variable
	The expected value of a numeric dependent variable

Question 5 (1 point)

A non-constant variance of the error terms in relation to the response variable is an issue involving:

Question 5 options:

	Heteroscedasticity
	Correlation with the error term
	The error terms are not normally distributed
	A non-linear relationship between the dependent and independent variables

Question 6 (1 point)

Logistic Regression models:

Question 6 options:

	The probability of outcome of the dependent variable
	The average of the dependent variable
	The log of the dependent variable
	The square root of the dependent variable

Question 7 (1 point)

In logistic regression, the logistic function is used to:

Question 7 options:

	Transform the variable using a log function
	Produce a bounded value between -1 and +1
	Produce a probability value
	Produce a value below 0 and above 1

Question 8 (1 point)

What is the commonly used estimator employed to estimate the parameters in logistic regression?

Question 8 options:

	Mean-Squared Error (MSE) Estimator
	Least-Square Estimator
	Maximum Likelihood Estimator (MLE)
	No estimator is used, there exists a closed-form solution

Question 9 (1 point)

Extending logistic regression to predict variables with multiple categories is called

Question 9 options:

	Binomial logistic regression
	Multiple logistic regression
	Multinomial logistic regression
	Binary Logistic regression

Question 10 (1 point)

Logistic regression is a specific type of a generalized linear model. What is the appropriate distribution type for the dependent variable in logistic regression when used as a genialized linear model?

Question 10 options:

	Gaussian
	Binomial
	Poisson
	Gamma

Question 11 (1 point)

Which one describes best about model validation and model testing?

Question 11 options:

	If the training data is small, there is no need to have an independent test set, the valuation step can be used to evaluate the model performance
	Once the validation step is done, the validation set is combined with the test set to assess the overall performance of the model
	Both model validation and testing are ways to measure the model performance, only one is needed.
	Validation is used to help tune the model using a validation set while testing is used to evaluate the model performance using an independent set

Question 12 (1 point)

What are acceptable ways to evaluate a classification model performance? Select all that apply.

Question 12 options:

	Use out of sample
	Use training set
	Use independent test set
	Use cross validation

Question 13 (1 point)

A significant reduction in the model accuracy using a test set compared to a training set is an indication of?

Question 13 options:

	Underfitting
	Complexity
	Overfitting
	Non-linearity

Question 14 (1 point)

When does a simple accuracy do a reasonably good job in assessing the performance of a classification model?

Question 14 options:

	For data with many categories
	For normally distributed data
	For imbalanced data
	For a balanced data

Question 15 (1 point)

Which is more important for a classification model, precision or recall?

Question 15 options:

	It depends on the application; in some, precision is more important than recall and vise versa
	Recall as it evaluate the detection capability of the model
	Precision as it measures the true classification capability of the model
	The both are equally important and should be weighted proportionally