Home
Blog
University of California Irvine Linear Regression and Prediction Questions

University of California Irvine Linear Regression and Prediction Questions

Daniel Kevins

0 comments

A Naive Bayes model is assessing the following sentence: Get meds from Canadian pharmacies discreetly shipped to your door no prescription needed! Which of the following statements is true?

	a.	The model will be rendered ineffective by the word Canadian.
	b.	The model will assess each of the words in the sentence independently (i.e. regardless of what other words are in the sentence).
	c.	The model will look for suspicious multiple-word strings, such as discreetly shipped and no prescription .
	d.	The model will make a classification estimate for this sentence as either spam or ham without factoring in the overall percentage of e-mails in the training set that are either spam or ham.

Which types of predictors are generally most useful to include when building a multiple linear regression model?

	a.	Predictors that are highly correlated with the outcome variable and with other predictors.
	b.	Predictors with no clear correlative relationship with either the outcome variable or with other predictors.
	c.	Predictors that are highly correlated with the outcome variable, but not with other predictors.
	d.	Predictors that are negatively correlated with the outcome variable and positively correlated with other predictors.

The chart below shows the summary information for a simple linear regression model with iris petal width as the outcome variable and iris petal length as the input variable, with both measured in centimeters. What iris petal width would this model predict when the petal length is 10cm?

chart for Quiz 2 question 14

	a.	9.271
	b.	3.794
	c.	-3.631
	d.	4.158

Imagine that AD699 has five new teams of students, as shown below. Each team was asked to rank lectures 1 through 5 on a scale from 1 to 10. A correlation table based on those rankings is shown below. Using the method for determining correlation distance that we use in AD699, which pair of teams will have the smallest correlation distance?

chart for Quiz 2 question 11

	a.	GOLF and HOTEL
	b.	INDIA and JULIET.
	c.	HOTEL and INDIA.
	d.	KILO and JULIET.

A survey was conducted in which BU students were asked about whether they walked or used Uber to get to class. 515 students were included in the survey. 152 of the students indicated that they sometimes walk, but never use Uber. 250 of the students sometimes use Uber. 205 of the students in the survey never walk to class. What is the probability that a randomly-selected student from this survey neither walks nor takes Uber?

	a.	0.398
	b.	0.781
	c.	0.602
	d.	0.219

You want to build a model using Naive Bayes, but some of your predictors are continuous numerical values. What can you do?

	a.	You can bin those values into different groups, and then treat each group as a factor.
	b.	You need to rebuild the model, with more emphasis placed on the variables.
	c.	The best thing to do here is to reframe the data as a table.
	d.	Nothing can be done here — you need to use a different algorithm.

An analyst is attempting to build a multiple regression model. Her model will include up to 7 possible independent variables, along with 1 dependent variable. She will need to explore the one-to-one relationships among the independent variables first, in order to reduce the risk of multicollinearity. How many total relationships among the independent variables does she need to explore?

	a.	21
	b.	28
	c.	42
	d.	7

、The Music Genome Project is based on what type of model?

	a.	Association rules.
	b.	Collaborative-based filtering.
	c.	Content-based filtering.
	d.	Exhaustive search.

In a k-nearest neighbors model with 4 pairs of binary dummies as predictors, how many binary dummy pairs should be used in the model?

	a.	Only one pair (if the data has not been normalized first).
	b.	16
	c.	3
	d.	4

Imagine that AD699 has five new teams of students, as shown below. Each team was asked to rank lectures 1 through 5 on a scale from 1 to 10. A correlation table based on those rankings is shown below. Using the method for determining correlation distance that we use in AD699, what is the correlation distance between team INDIA and team GOLF?

chart for Quiz 2 question 10

	a.	1.492
	b.	-0.492
	c.	0.509
	d.	-1.492

Imagine that AD699 has five new teams of students, as shown below. Each team was asked to rank lectures 1 through 5 on a scale from 1 to 10. A correlation table based on those rankings is shown below. Using the method for determining correlation distance that we use in AD699, what is the correlation distance between team INDIA and team GOLF?

chart for Quiz 2 question 10

	a.	1.492
	b.	-0.492
	c.	0.509
	d.	-1.492

Which of the following statements about tree-based models is NOT true?

	a.	Tree-based models can handle missing data well (i.e. without degrading the results of the model).
	b.	Tree-based models are especially good at identifying the relationships among predictors.
	c.	Tree-based models can handle the presence of outliers well (i.e. without degrading the results of the model).
	d.	In order to work well as classifiers, tree-based models require large training data sets.

In which of these models does an analyst start by including ALL of the possible independent variables, and then eliminate some of those variables, one at a time?

	a.	Stepwise regression.
	b.	Forward selection.
	c.	Ordinary least squares.
	d.	Backward elimination.

TRUE or FALSE: If Rectangle A has a higher Gini index than Rectangle B, then we can expect it to be less homogenous.

True

False

TRUE or FALSE: A hamming distance can be negative.

True

False

About the Author

Daniel Kevins

Follow me