ISYE6501 ISYE 6501 Midterm

ISYE6501 / ISYE 6501 Midterm Exam 1 (Latest Update 2025 / 2026) Intro to Analytics Modeling | Questions & Answers | Grade A | 100% Correct - Georgia Tech

Question:

what is T-statistic?

Answer:

the Coefficient divided by its standard error. It is related to the P-value.

Question:

True or False: R-Squared values of 0.4 or 0.3 are quite good?

Answer:

TRUE

• / 4

Question:

what are common outputs you will see when you call a regression model?these help you determine the importance of coefficients.

Answer:

P-value, T-Statistic, R2, CI, and the coefficients.

Question:

what is considered better: a lower P value, or a higher p value?

Answer:

lower. generally, anything under 0.05 (5%)

Question:

if an R-Squared value is 59%, what does that represent. inversely, what does the other 41% represent.?

Answer:

59% accounts for the variability in the data.41% represents randomness or other factors.

• / 4

Question:

why would you want to add a quadratic term in a regression model? what does it do?

Answer:

to transform the data to make the fit linear (allowing you to use a linear model) it creates a new column for the data.

Question:

What is BoxCox? why use it?

Answer:

a logarithmic transformation that stretches out the smaller range to enlarge its variability, and shrinks the larger range to reduce its variability. You use it to deal with Unequal Variances (Heteroscedasticity)

use QQ plot to determine if you NEED to do this.

Question:

what is Heteroscedasticity?

Answer:

unequal variance. The Variance is different in different ranges of the data.

• / 4

Question:

when should you detrend a model?

Answer:

when your going to use a factor based model like time series. Think of inflation and how it makes a big difference in a model outcome.

Question:

How might you detrend?

Answer:

Factor-by-Factor: fit a one dimensional regression to it (like linear: Y = a0 +a1x)

Question:

Which factor is most important for classification? Factors on horizontal or vertical axis?

Answer:

almost all that matters is whether the vertical axis value is above or below a specific value.

• / 4