STAT 432
CBTF Exam III Policies
Date and Time
- Start: Wednesday, December 4
- End: Friday, December 6
Please see the CBTF scheduler:
CBTF Policies
Please be aware of the CBTF policies:
If you encounter a technology issue during the exam, you must file an incident report with a proctor. Without documentation, we cannot provide accommodation for technology errors encountered during an exam.
Many technology issues can be solved by simply restarting RStudio.
Academic Integrity
In short, do not cheat. In addition to the Academic Integrity policies outlined by the CBTF, STAT 432 will hold itself to a higher standard. Any violation will be punished as harshly as possible, which at minimum is a failing grade on the exam.
A non-exhaustive list of academic integrity violations:
- Informing another student of your exam date and time.
- Informing another student about exam material, generally or specifically.
You are correct that some of these things may be hard to detect, but in addition to writing the exam in a manner that makes transferring this type of information less useful, we have other methods that you are likely unaware of to detect the above violations.
Be aware that the exam window goes until Friday. However, this does not mean that you may discuss the exam at that time as there may be makeup exams in rare circumstances. Do not discuss any part of the exam until I announce that it is OK to do so.
Content
The exam will have 10 questions. The following list outlines what these questions will be.
- Cross-Validation “The Long Way” (Any model. Any metric.)
- A repeat of the Exam II SVM question.
- Another “new” classifier like Exam II.
- Tuning a classifier with caret.
- Tuning regression with caret.
- A probability question.
- Using
randomForest()
for classification.
- Using
randomForest()
for regression.
- Using
cv.glmnet()
for classification.
- Using
cv.glmnet()
for regression.
Provided Notes
\[\text{Sensitivity} = \frac{\text{TP}}{\text{P}} = \frac{\text{TP}}{\text{TP + FN}}\] \[\text{Specificity} = \frac{\text{TN}}{\text{N}} = \frac{\text{TN}}{\text{TN + FP}}\]
These, and perhaps other similar definitions (related to binary classification) will be provided with any relevant problems. No additional materials will be provided.
FAQ
How many problems are on the exam?
How long will it take to do the exam?
- If you are well prepared and manage your time efficiently: Under 1 hour and 50 minutes.
- If you are ill prepared or manage your time poorly: Over 1 hour and 50 minutes.
- You should not focus on this. Focus on practicing and being able to do questions of the types you’ve seen on quizzes. Let me worry about this. And believe me, I worry about this a lot.
Will the exam be curved?
Home