Blog

Question 5 Prediction (8 points) Use testData for this quest…

Question 5 Prediction (8 points) Use testData for this question a) (4 points) Using testData, predict the Savings_Rate_Percent with both model1 and model2.i) Show the predictions of both models along with the true values.ii) Calculate the mean squared prediction error (MSPE) of each model. Which model predicts better on the test data? b) (4 points) Consider a new individual who owns their home (Housing_Type = “Own”), has monthly expenses of 4000 USD and a credit score of 700. Using model1:i) Compute the 95% confidence interval for the mean response and the 95% prediction interval for this new individual.ii) Provide an interpretation of each interval in the context of the problem. Why is the prediction interval wider than the confidence interval?

Question 3 Model Diagnostics and Transformation (19 points)…

Question 3 Model Diagnostics and Transformation (19 points) Use trainData for this question a) (4 points) Perform the following model diagnostics on model2 (the full model created in Question 2b):i) Check for the linearity assumption.ii) Check for constant variance.                                                                                                                                                          iii) Check for normality. Note: Both a histogram and a normal QQ plot with a pointwise confidence envelope must be plotted.Explain your conclusions. b) (2 points) Based on your conclusions in Q3a, would you propose a transformation of the response variable? Explain with reasoning. c) (2 points) i) Create a linear regression model, named model4, that uses the log-transformed Savings_Rate_Percent as the response and all the predictors in trainData. Display the summary.(2 points) ii) Compare the R-squared values of model2 and model4. Did the transformation improve the explanatory power of the model? d) (3 points) Calculate the VIF of each predictor in model2. Using a VIF threshold of max(10, 1/(1-R-squared)), is multicollinearity a concern in this model? e) (3 points) Create a plot of the Cook’s distances for model2. Using a threshold of 4/n, how many outliers are identified? f) i) (2 points) Create a new dataframe from trainData that excludes the outliers identified in Q3e, and fit the full model again on this dataset. Call it model_outlier. Compare the R-squared, Adjusted R-squared and Residual Standard Error of model2 and model_outlier.ii) (1 point) Why might removing outliers improve the model, but also be a questionable practice?

Midterm Exam – Part 2 Instructions The R/Python Jupyter N…

Midterm Exam – Part 2 Instructions The R/Python Jupyter Notebook files include the questions , the empty code chunk sections for your code, and the text blocks for your responses. Answer the questions below by completing the R/Python Jupyter Notebook file. You may make slight adjustments to get the file to convert but otherwise keep the formatting the same.   Once you’ve finished answering the questions, submit your exam in pdf format, BOTH to Canvas and to Gradescope. Please use the Gradescope link for the submission. The Gradescope link expires after 10 minutes, so make sure you submit PDF to Gradescope within 10 minutes! Resubmission within the 10-minute window is allowed.  Please make your submission within the exam window as there are penalties for late submissions.   To maintain the integrity of this course: 1. Do not plagiarize (even if it is a particular question).  2. Do not use any AI tools such as chatGPT or CoPilot. Students violating the Honor Code will be reported to Georgia Tech’s Office of Student Integrity. Ready? Let’s begin…   Data Set: Personal_Financial_Wellness.csv Starter Templates: Summer2026_midterm_R_Starter_template.ipynb Summer2026_midterm_Python_Starter_template.ipynb