ISYE 6501 - MIDTERM 2 2026 UPDATE | QUESTIONS AND ANSWERS |
ss ss ss ss ss ss ss ss ss ss ss
WITH COMPLETE SOLUTIONS ss ss ss
Terms in this set (160)
ss ss ss ss
when the # of factors is close to or larger than the # of
ss ss ss ss ss ss ss ss ss ss ss ss ss
when might overfitting occur
ss ss ss ss data points causing the model to potentially fit
ss ss ss ss ss ss ss
ss too closely to random effects
ss ss ss ss
Why are simple models better
ss ss ss ss less data is required; less chance of insignificant
ss ss ss ss ss ss ss
ss than complex ones
ss ss ss factors and easier to interpret
ss ss ss ss
we select the best new factor and see if it's good
ss ss ss ss ss ss ss ss ss ss
ss enough (R^2, AIC, or p-value) add it to our
ss ss ss ss ss ss ss ss
model and fit the model with the current set of
ss ss ss ss ss ss ss ss ss ss
factors.
ss
what is forward selection
ss ss ss
Then at the end we remove factors that are lower
ss ss ss ss ss ss ss ss ss
ss than a certain threshold
ss ss ss
1/26
, we start with all factors and find the worst on a
ss ss ss ss ss ss ss ss ss ss
ss supplied threshold (p = 0.15). If it is worse we
ss ss ss ss ss ss ss ss ss
ss remove it and start the process over. We do that
ss ss ss ss ss ss ss ss ss
ss until we have the number of factors that we want
ss ss ss ss ss ss ss ss ss
ss and then we move the factors lower than a
ss ss ss ss ss ss ss ss
what is backward elimination
ss ss ss
ss second threshold (p = .05) and fit the model with
ss ss ss ss ss ss ss ss ss
ss all set of factors
ss ss ss
2/26
, ISYE 6501 - Midterm 2ss ss ss ss
Study
backward elimination. We can either start withss ss ss ss ss ss
all
factors or no factors and at each step we remove
ss ss ss ss ss ss ss ss ss
ss or add a factor. As we go through the procedure
ss ss ss ss ss ss ss ss ss
ss after adding each new factor and at the end we
ss ss ss ss ss ss ss ss ss
what is stepwise regression
eliminate right away factors that no longer appear.
ss ss ss
ss ss ss ss ss ss ss s
what type of algorithms are
ss ss ss ss Greedy algorithms - at each step they take one
ss ss ss ss ss ss ss ss
stepwise selection?
ss ss ss thing that looks best ss ss ss
a variable selection method where the coefficients
ss ss ss ss ss ss
ss are determined by both minimizing the squared
ss ss ss ss ss ss
ss error and the sum of their absolute value not being
ss ss ss ss ss ss ss ss ss
ss over a certain threshold t
ss ss ss ss
what is LASSO ss ss
How do you choose t
ss ss ss ss use the lasso approach with different values of t and
ss ss ss ss ss ss ss ss ss
ss in LASSO
ss see which gives the best trade off
ss ss ss ss ss ss ss
why do we have to scale
ss ss ss ss ss if we don't the measure of the data will artificially
ss ss ss ss ss ss ss ss ss
ss the data for LASSO
ss ss ss ss affect how big the coefficients need to be
ss ss ss ss ss ss ss
3/26
ss ss ss ss ss ss ss ss ss ss ss
WITH COMPLETE SOLUTIONS ss ss ss
Terms in this set (160)
ss ss ss ss
when the # of factors is close to or larger than the # of
ss ss ss ss ss ss ss ss ss ss ss ss ss
when might overfitting occur
ss ss ss ss data points causing the model to potentially fit
ss ss ss ss ss ss ss
ss too closely to random effects
ss ss ss ss
Why are simple models better
ss ss ss ss less data is required; less chance of insignificant
ss ss ss ss ss ss ss
ss than complex ones
ss ss ss factors and easier to interpret
ss ss ss ss
we select the best new factor and see if it's good
ss ss ss ss ss ss ss ss ss ss
ss enough (R^2, AIC, or p-value) add it to our
ss ss ss ss ss ss ss ss
model and fit the model with the current set of
ss ss ss ss ss ss ss ss ss ss
factors.
ss
what is forward selection
ss ss ss
Then at the end we remove factors that are lower
ss ss ss ss ss ss ss ss ss
ss than a certain threshold
ss ss ss
1/26
, we start with all factors and find the worst on a
ss ss ss ss ss ss ss ss ss ss
ss supplied threshold (p = 0.15). If it is worse we
ss ss ss ss ss ss ss ss ss
ss remove it and start the process over. We do that
ss ss ss ss ss ss ss ss ss
ss until we have the number of factors that we want
ss ss ss ss ss ss ss ss ss
ss and then we move the factors lower than a
ss ss ss ss ss ss ss ss
what is backward elimination
ss ss ss
ss second threshold (p = .05) and fit the model with
ss ss ss ss ss ss ss ss ss
ss all set of factors
ss ss ss
2/26
, ISYE 6501 - Midterm 2ss ss ss ss
Study
backward elimination. We can either start withss ss ss ss ss ss
all
factors or no factors and at each step we remove
ss ss ss ss ss ss ss ss ss
ss or add a factor. As we go through the procedure
ss ss ss ss ss ss ss ss ss
ss after adding each new factor and at the end we
ss ss ss ss ss ss ss ss ss
what is stepwise regression
eliminate right away factors that no longer appear.
ss ss ss
ss ss ss ss ss ss ss s
what type of algorithms are
ss ss ss ss Greedy algorithms - at each step they take one
ss ss ss ss ss ss ss ss
stepwise selection?
ss ss ss thing that looks best ss ss ss
a variable selection method where the coefficients
ss ss ss ss ss ss
ss are determined by both minimizing the squared
ss ss ss ss ss ss
ss error and the sum of their absolute value not being
ss ss ss ss ss ss ss ss ss
ss over a certain threshold t
ss ss ss ss
what is LASSO ss ss
How do you choose t
ss ss ss ss use the lasso approach with different values of t and
ss ss ss ss ss ss ss ss ss
ss in LASSO
ss see which gives the best trade off
ss ss ss ss ss ss ss
why do we have to scale
ss ss ss ss ss if we don't the measure of the data will artificially
ss ss ss ss ss ss ss ss ss
ss the data for LASSO
ss ss ss ss affect how big the coefficients need to be
ss ss ss ss ss ss ss
3/26
