Predictive Modeling and Analytics Coursera Quiz Answers

All Weeks Predictive Modeling and Analytics Coursera Quiz Answers

Welcome to the second course in the Data Analytics for Business specialization!

This course will introduce you to some of the most widely used predictive modeling techniques and their core principles. By taking this course, you will form a solid foundation of predictive analytics, which refers to tools and techniques for building statistical or machine learning models to make predictions based on data.

Enroll in Predictive Modeling and Analytics on Coursera

Predictive Modeling and Analytics Week 01 Quiz Answers

Quiz 1: Predictive Modeling and Analytics

Q1. At what stage(s) of Data Exploration would you address missing values in a data set?

  • Data transformation
  • Data clean-up
  • Data reduction

Q2. Which of the following statements regarding data transformation and data reduction is correct?

  • Data transformations work on individual variables, while data reduction works on a set of variables
  • Only data transformation would create dummy variables
  • The goal of data transformations is to create larger datasets while the goal of data reduction is to create smaller datasets
  • Data transformations are out of style; data reduction is the modern man’s tool

Q3. What does a data value measure after centering and scaling has been applied?

  • Accuracy
  • The number of standard deviations between each data point and the median
  • The number of standard deviations between each data point and the mean
  • Slope

Q4. Why would one want to center and scale a set of data?

  • So multiple variables in the data set are on a common scale
  • To make all data values positive
  • To remove duplicates
  • To make data easier to interpret

Q5. For the following three questions, match the Box-Cox Transformation associated with the given value of lambda.

When Lambda = 0, transformation is

  • Logarithmic
  • Cubed polynomial
  • Inverse
  • Square root

Q6. When Lambda = 0.5, transformation is

  • Logarithmic
  • Cubed polynomial
  • Inverse
  • Square root

Q7. When Lambda = – 1, transformation is

  • Logarithmic
  • Cubed polynomial
  • Inverse
  • Square root

Q8. What is the purpose of applying a Data Reduction?

  • To generate a larger set of variables
  • To make all variables positively valued
  • To use a smaller set of variables to capture most of the information in the original variables

Q9. What must be done to variables of a data set before applying principal component analysis and why?

  • You must scale the variables so that only outliers are considered as principal components
  • You must scale the variables so that principal components are not dominated by variables of much larger scale
  • You must make all variables negative to work with values of the same sign
  • You must take the square root of all data values to reduce the overall magnitudes of the data set

Q10. Which of the following can be an appropriate way to deal with missing values? (Select all that apply.)

  • Removing the columns or rows with missing values.
  • Imputing a value with averages of all other records.
  • Imputing a value from “similar” data points.
  • Making “missing” its own category.

Q11. Your organization asks you to analyze a dataset that shows the number of FreeFly ALTA drones sold in 2016. You noticed that only 2 drones were sold the day after Black Friday, while the average number of drones sold in 2016 is around 100 a day. What is the most probable explanation for this small data value?

  • It’s a missing value that someone filled in with a guess
  • There was a glitch in the system and the data value was corrupted
  • It’s a censored value that was inputted incorrectly
  • It’s a censored value; drone inventory probably ran out

Q12. What are the risks of replacing a missing value with a guess? (Select all that apply.)

  • None, the database is capable of correcting input mistakes
  • Introducing biases
  • Distorting the data set
  • Falsifying results

Q13. Why removing all data records with missing values is often not a good way to deal with missing values? (Select all that apply.)

  • Some modeling tools require a data value for each row/column.
  • A dataset is incomplete if there are missing values.
  • We may end up with too little data to conduct meaningful analysis.
  • The pattern of missing values can have high predictive power.

Q14. The table below shows the number of patients visiting a clinic on ten consecutive business days.

DateWeekdayPatient Count
6/23/14Monday16
6/24/14Tuesday12
6/25/14Wednesday12
6/26/14Thursday14
6/27/14Friday11
6/30/14Monday12
7/1/14Tuesday14
7/2/14Wednesday11
7/3/14Thursday15
  • 0
  • 13
  • 11
  • 15.7

Q15. What are the characteristics of an outlier? (Select all that apply.)

  • It is the data point most proximal to the mean
  • It is the pivot point for the overall pattern that the data follows
  • It falls far outside the overall data pattern
  • It is above or below 3 standard deviations of the mean

Q16. A data point is not considered an outlier unless it deviates dramatically on either the x-axis or the y-axis?

  • True
  • False

Q17. Why do outliers exist? (Select all that apply.)

  • Data recording errors
  • Legitimate but odd observations
  • Entropy of a system
  • Distortion of time

Q18. Which statistical measure is more resistant to outliers?

  • Mean
  • Median
  • Standard deviation
  • Range

Q19. To say a variable is degenerate means which of the following: (Select all that apply.)

  • The variable is immoral and corrupt
  • The variable can only take on a single value
  • When plotted, the variable is modeled with an exponential decay
  • The variable is a zero variance variable

Q20. Which of the following is a remedy to collinearity issues in regression analysis?

  • Adding more dummy variables
  • Cut the data set in half
  • Removing zero variance and near zero variance variables
  • Duplicate the data set

Q21. Check all valid observations about the set of box plots below: (Select all that apply.)

  • The minimum of Box 4 is less than the minimum of Box 5
  • The median of Box 1 is greater than the median of all other boxes
  • Box 3 has a larger median than Box 4
  • Box 3 has the largest range of all boxes

Q22. Which function on the graph is both linear and displays a positive relationship?

  • 1
  • 2
  • 3
  • 4
  • 5

Q23. How could this time series graph be improved for data visualization purposes?

  • Swap the x and y axis
  • Adding a line to highlight trend
  • Convert it to a pie chart
  • Shade the background with a visually appealing texturized color

Q24. Given the table below, how many dummy variables should you create based on the number of categories present?

  • 1
  • 2
  • 3
  • 4

Predictive Modeling and Analytics Week 02 Quiz Answers

Quiz 1 answers

Q1. Which type of target variable are we dealing with in linear regression?

  • Binary
  • Categorical
  • Continuous
  • Imaginary

Q2. We cannot perform linear regression unless both the target variable and predictor variables are continuous.

  • True
  • False

Q3. Given the graph below, we see a red best fit line has been applied to map the trend of the data. We also see specific clusters of points have been circled out and labeled 1, 2, and 3. Which cluster has the greatest residual value magnitude ( ignore positive and negative signs, just compare numeric values ) ?

  • 1
  • 2
  • 3

Q4. What is the validation set used for in predictive modeling?

  • To fit the models
  • To evaluate the various models
  • To increase the size of our training set
  • To average the training set data

Q5. Calculate the sum of squared errors for the validation data set given below. Recall, you square the residual values before summing them .

  • 18.2
  • 248
  • 10
  • 439

Q6. Why can multicollinearity cause problems in multiple regression? (Select all that apply.)

  • It creates unstable estimate
  • It creates problem in model interpretation
  • You cannot make predictions based on regression models with multicollinearity issues
  • It makes estimating the model impossible

Q7. How many transformed variables can we create based on one predictor variable?

  • An unlimited number
  • 2
  • None
  • 1

Q8. The figures here show the scatter plot with different data transformations. The top left figure is the scatter plot of list price vs SQFT when no transformation is performed. The top right one shows the scatter plot when a log transformation is applied to SQFT. The bottom left figure shows the plot where a log transformation is applied to list price, while the bottom right figure shows the plot where log transformation is applied to both variables.

Which one gives us better model fit?

  • Top right figure
  • Top left figure
  • Bottom left figure
  • Bottom right figure

Q9. What do you achieve when you apply a log transformation to a variable in your data set?

  • It makes highly skewed distributions less skewed
  • It compresses data to make big sets more manageable
  • It removes negative data values
  • It removes duplicate values

Predictive Modeling and Analytics Week 03 Quiz Answers

Quiz 2: Application Assignment

Q1. In this assignment we are continuing to work with customer reward programs (review assignments from Week 1 if you haven’t completed them). The data is in the file

crp_cleandata
XLSX File
Download file
In this exercise, you will complete a predictive modeling task where the target variable is continuous based on the data in the shared file. First remove all rows where either the Reward or NumStores column takes the value 0. Also remove all rows where the rewards do not expire (ExpirationMonth=999). [Hint: You can sort the relevant columns to quickly find the rows to delete. ] How many rows are left after deleting these irrelevant rows, not counting the header row? What is the sum of the ExpirationMonth column?

  • 48, 1335
  • 48, 336
  • 46, 1335
  • 46, 336

Q2. Consider linear regression models with ExpirationMonth column as the target variable. Find the model with one predictor variable and the highest R-squared. Consider the following set of predictor variables: Salerank, X2013USSales, X2013WorldSales, NumStores,RewardSize, and ProfitMargin. Which variable did you choose?

  • X2013USSales
  • X2013WorldSales
  • Salerank
  • NumStores

Q3. What is the estimated intercept coefficient of the model?

  • 5.7082
  • 4.8285
  • 0.8898
  • 0.2537

Q4. What is the estimated slope coefficient of the model?

  • 34.9427
  • 4.8285
  • 0.8898
  • 0.2537

Q5. Data transformation is a great way to improve model fit. Now consider the log transformation for the model identified in the previous question. [Hint: Use log function to create the transformed columns.] You can choose to transform neither of them, one of them, or both of them. You should have four different models.

  • Model 1: neither variable is transformed; this gives you the same model as in the previous question.
  • Model 2: only the target variable is transformed
  • Model 3: only the explanatory variable is transformed
  • Model 4: both variables are transformed.

Report the R-squared values of all four models.

What is the R-squared for Model 1?

  • 4.8285
  • 0.8898
  • 0.2537
  • 6.6175

Q6. R-squared for Model 2 is ( report answer using 4 decimal places i.e. x.xxxx ):

0.0704

Q7. R-squared for Model 3 is ( report answer using 4 decimal places i.e. x.xxxx ):

0.1446

Q8. R-squared for Model 4 is ( report answer using 4 decimal places i.e. x.xxxx ):

0.0652

Q9. Which model gives the best fit based on the R-squared value?

  • Model 1
  • Model 2
  • Model 3
  • Model 4

Q10. Our analysis so far shows that variable transformation does not improve the model fit. Another way to improve model fit is to add more explanatory variables on the right side. Again consider the following set of predictor variables: Salerank, X2013USSales, X2013WorldSales, NumStores, RewardSize, and ProfitMargin. Add one more variable to the best model you identified in the previous question. Which variable will you add? Hint: The correct additional variable gives the highest R-squared value.

  • RewardSize
  • X2013USSales
  • ProfitMargin
  • Salerank

Q11. What is the R-squared for the model with the additional variable added ( report answer using 4 decimal places i.e. x.xxxx )?

Enter answer here

Q12. One way to figure out whether a linear regression model explains a particular data point well is to look at the residual. For which retailer do you have the highest absolute value of residual based on your result in the previous question?

  • Macy’s
  • Whole Foods
  • TJX
  • Starbucks

Q13. For which retailer do you have the lowest residual based on your result in the previous question?

  • Macy’s
  • Whole Foods
  • TJX
  • Starbucks

Predictive Modeling and Analytics Week 04 Quiz Answers

Quiz 1

Q1. A soccer team is believed to have a 8 to 2 odds of winning the election. What is the probability of winning for the candidate?

  • 0.2
  • 0.25
  • 0.8
  • 2

Q2. It is estimated that an appointment with a 10 day lag for a male patient has a predicted probability of 0.1372 of cancelling. Compare this with the predicted cancellation probability for a female patient who also has an appointment with a 10 day lag.

Assume that value of gender variable is 1 for male patients and 0 for females. Also, assume that the estimated coefficient for gender is -0.3572 , beta-0 is -1.6515 , beta-1 is .01699.

  • A female is less likely to cancel by 2.4%
  • A female is more likely to cancel by 4.8%
  • A female is equally likely to cancel as a male
  • A female is more likely to cancel by 6.9%

Q3. Given the following table, does the shaded quadrant represent true positives, true negatives, false positives, or false negatives? Assume cancellation is denoted by 1 and arrival is denoted by 0.

  • True positives
  • True negatives
  • False positives
  • False negatives

Q4. Answer Question 4 and Question 5 based on this confusion matrix:

It is believed that we can reverse 60% of cancellations with reminder phone calls. We decided to place reminder calls for all appointments that are predicted to cancel. For how many cases can we reverse cancellation (round to the nearest integer) ?

Assume cancellation is denoted by 1 and arrival is denoted by 0.

Enter answer here

Q5. Using the confusion matrix and answer found in Question 4….

Assume the cost of placing a reminder call is $1 using an automated system and the benefit of serving a patient is $50. What is the profit of placing reminder calls for all appointments predicted to cancel (round to the nearest integer)?

  • 2300
  • 3300
  • 4300
  • 5300

Q6. Consider the following confusion matrix when answering the next two questions.

What is the precision?

  • 0.8571
  • 0.5294
  • 36
  • 0.3462

Q7. What is the accuracy of the model?

  • 0.1290
  • 0.3462
  • 0.6261
  • 72

Quiz 2: Application Assignment

Q1. Let’s reconsider the customer reward program dataset. In this exercise, you will complete a predictive modeling task where the target variable is binary. Using the following data file for this exercise:

crp_cleandata
XLSX File
Download file
The dataset also contains a column IndustryType, which is created based on the column Industry in the raw data. Note that Industry has many categories. The analyst who prepared the data chose to combine some categories, which resulted in the column IndustryType. IndustryType has five categories: Department, Discount, Grocery, Restaurants, Specialty. You can create a set of dummy variables based on IndustryType in XLMiner by using the Transform functions.

Part I.

Consider logistic regression models with Reward column as the target variable. Fit the model with two indicator variables, one indicating whether a retailer is a discount store (i.e., IndustryType is Discount), and the other indicating whether a retailer is a grocery store (i.e., IndustryType is Grocery). Report the coefficient estimates in the next three questions. [Hint: After you create the dummy variables, use them as Selected Variables (instead of Categorical Variables) in the first step of Logistic Regression.]

What is the estimated intercept coefficient?

  • 10
  • 0.5108
  • 133.47
  • 0.03023

Q2. What is the estimated coefficient for IndustryType_Discount (round the answer to 4 decimal places i.e. x.xxxx ) ?

  • -0.9627

Q3. What is the estimated coefficient for IndustryType_Grocery (round the answer to four decimal places i.e. x.xxxx) ?

  • -0.7339


Q4. What is the number of true positives? (Specify a whole number.)

  • 40


Q5. What is the number of true negatives? (Specify a whole number.)

  • 21

Q6. Part II.

Split the dataset into training and validation sets using a 60:40 split (set the seed for partitioning to 12345; this should be the default value if you have not changed it). [Hint: note that there two Partition buttons in XLMiner ribbon. You should use the Partition->Standard Partition in the Data Mining group.] Report the new coefficient estimates in the next three questions. Use the same two predictor variables as in Part I.

What is the estimated intercept coefficient (round the answer to 4 decimal places i.e. x.xxxx) ?

  • Enter answer here

Q7. What is the estimated coefficient for IndustryType_Discount (round the answer up to 4 decimal places i.e. x.xxxx ) ?

  • Enter answer here

Q8. What is the estimated coefficient for IndustryType_Grocery (round the answer to 4 decimal places i.e. x.xxxx ) ?

  • Enter answer here

Q9. How many observations are in the training set?

  • 60

Q10. What is the number of true positives on the validation data? (Specify a whole number.)

  • Enter answer here

Q11. What is the number of true negatives on the validation data? (Specify a whole number.)

  • Enter answer here

Q12. (Part 3) By default, XLMiner uses the cutoff threshold 0.5. Repeat Part II with a cutoff threshold 0.3. What are the numbers of true positives and true negatives on the validation data?

Report the number of true positives:

  • Enter answer here

Q13. Report the number of true negatives:

  • 0

Week 4:

Q1. Consider the following split in the appointment data.

What is the Gini index for the branch Age<65.5?

  • 0.25
  • 0.375
  • 0.50
  • 0.875

Q2. The bagging procedure can reduce the variance of a predictive model. Check all true statements about the bagging method. (Check all that apply.)

  • Helps avoid overfitting of the dataset
  • Helps group similar data outliers
  • Has access to multiple training sets
  • Can be applied to tree models

Q3. What do the bagging and random forest methods have in common?

  • Both methods grow multiple numbers of trees
  • Both methods operate on only 2 trees
  • Both methods sample the validation set
  • Both methods increase the variance of a dataset

Q4. What sets the random forest algorithm apart from bagging and boosting algorithms?

  • It operates on bootstrap sets
  • It focuses on reducing correlation among models
  • It involves multiple tree models
  • It predicts the average variance of a set

Q5. What relationship do the hidden layer nodes have to the output layer nodes based on the diagram?

  • One to one
  • One to many
  • Many to many
  • Many to one

Q6. True or False: both linear regression and logistic regression can be viewed as a neural network with no hidden layers.

  • True
  • False

Quiz 2

Q1. Let’s once again consider the customer reward program dataset. For your convenience, here the original data set.

data-download-wk4
XLSX File
Download file
In this exercise, we use machine learning methods including trees and neural networks. Set all random seeds to 12345 in XLMiner; this should be the default value if you have not changed them. Also use default options in XLMiner unless directed otherwise.
Q1. Build a full classification tree model with Reward column as the target variable and the following set of predictor variables: Salerank, X2013USSales, X2013WorldSales, ProfitMargin, NumStores, and IndustryType. Which feature has the highest feature importance?

[Hint: Use Classify in the Data Mining ribbon since Reward is a 0-1 variable.]

  • X2013USSales
  • IndustryType
  • ProfitMargin
  • NumStores
  • X2013WorldSales
  • Salerank

Q2. Split the dataset into training and validation sets using a 60:40 split. Use the same columns as in the previous question.

What is the precision of the validation data for a bagged tree (round the answer to 4 decimal places, i.e. x.xxxx)?

  • Enter answer here

Q3. What is the precision of the validation data for the boosted tree (round the answer to 4 decimal places, i.e. x.xxxx)?

  • Enter answer here

Q4. Build a neural network model with RewardSize column as the target variable and the following set of predictor variables: Salerank, X2013USSales, X2013WorldSales, NumStores, and profit margin. Note that the RewardSize variable is only relevant for retailers that use reward programs. For your convenience, here is the data file with values for RewardSize, which should be used for model building.

Split the dataset into a training and validation set using a 60:40 split. Also, remember to scale the data. Report RMSE on the validation data for the following models: (i) bagged neural net, (ii) boosted neural net. [Hint: Use Predict in the Data Mining ribbon since RewardSize is a continuous variable.]

Report RMSE on the validation data for the bagged neural net (round the answer to 4 decimal places, i.e. x.xxxx)?

  • 8.4296

Q5. Report RMSE on the validation data for the boosted neural net (round the answer to 4 decimal places, i.e. x.xxxx):

  • 9.8768
Predictive Modeling and Analytics Coursera Course Review:

In our experience, we suggest you enroll in the Predictive Modeling and Analytics Coursera and gain some new skills from Professionals completely free and we assure you will be worth it.

Predictive Modeling and Analytics Coursera course is available on Coursera for free, if you are stuck anywhere between quiz or graded assessment quiz, just visit Networking Funda to get Predictive Modeling and Analytics Coursera Quiz Answers.

This Course is a part Advanced Business Analytics Specialization

Conclusion:

I hope this Predictive Modeling and Analytics Coursera Quiz Answers would be useful for you to learn something new from this Course. If it helped you then don’t forget to bookmark our site for more Coursera Quiz Answers.

This course is intended for audiences of all experiences who are interested in learning about Data Science in a business context; there are no prerequisite courses.

Keep Learning!

Get all Course Quiz Answers of Advanced Business Analytics Specialization

Introduction to Data Analytics for Business Coursera Quiz Answers

Predictive Modeling and Analytics Coursera Quiz Answers

Business Analytics for Decision Making Coursera Quiz Answers

Communicating Business Analytics Results Coursera Quiz Answers

Leave a Reply

error: Content is protected !!