MISY 262 FINAL EXAM COMPREHENSIVE EXAM
QUESTIONS WITH DETAILED VERIFIED AND 100%
ACCURATE ANSWERS BRAND NEW EXAM ALREADY
GRADED A+ PASS
Statistical Significance Ans✓✓✓We have enough evidence that the
coefficient is different than zero and can be included in the model.
Nominal Ans✓✓✓Categorical variables with more than 2 category
values, but no logical order and no numerical values associated
Ordinal Ans✓✓✓Same as nominal, but with logical order
Error Ans✓✓✓Actual - Predicted
Multicollinearity Ans✓✓✓- Exists when there is a strong/medium
correlation between two or more predictors in a multiple regression
model.
- Two variables are totally INDEPENDENT if their correlation is 0.
- Stronger the correlation, the more alike 2 variables are (contribution
less valuable)
The less collinearity the more directly the coefficients can be interpreted
Perfect Multicollinearity Ans✓✓✓- Correlation between two or more
variables -1 or 1
- It exists when one variable is a linear combination of another variable
, - R will not run and show error
Partial Multicollearnity Ans✓✓✓- Correlation between two or more
variables is >0.25 or <-0.25
- Model is still useable and can make predictions
- R^2 is still interpretable but may not increase by much
Coefficients of variables exhibiting multicollinearity become difficult to
interpret
Checking for Multicollinearity Ans✓✓✓>
cor(cbind(over=d$pctoversized, damage=d$pctdamaged,
partner=d$partnerdelivery))
Good Linear Model (Assumption 1 Linear Model) Ans✓✓✓First:
"Good"
- There are no major independent variables we are leaving out
- There are no independent variables included that do not belong
Second: "Linear"
- This is where scatter plots are useful
- Check that each independent variable (predictor) exhibits a linear
relationship with the response
- Straight line
- plot(dependent, independent)
QUESTIONS WITH DETAILED VERIFIED AND 100%
ACCURATE ANSWERS BRAND NEW EXAM ALREADY
GRADED A+ PASS
Statistical Significance Ans✓✓✓We have enough evidence that the
coefficient is different than zero and can be included in the model.
Nominal Ans✓✓✓Categorical variables with more than 2 category
values, but no logical order and no numerical values associated
Ordinal Ans✓✓✓Same as nominal, but with logical order
Error Ans✓✓✓Actual - Predicted
Multicollinearity Ans✓✓✓- Exists when there is a strong/medium
correlation between two or more predictors in a multiple regression
model.
- Two variables are totally INDEPENDENT if their correlation is 0.
- Stronger the correlation, the more alike 2 variables are (contribution
less valuable)
The less collinearity the more directly the coefficients can be interpreted
Perfect Multicollinearity Ans✓✓✓- Correlation between two or more
variables -1 or 1
- It exists when one variable is a linear combination of another variable
, - R will not run and show error
Partial Multicollearnity Ans✓✓✓- Correlation between two or more
variables is >0.25 or <-0.25
- Model is still useable and can make predictions
- R^2 is still interpretable but may not increase by much
Coefficients of variables exhibiting multicollinearity become difficult to
interpret
Checking for Multicollinearity Ans✓✓✓>
cor(cbind(over=d$pctoversized, damage=d$pctdamaged,
partner=d$partnerdelivery))
Good Linear Model (Assumption 1 Linear Model) Ans✓✓✓First:
"Good"
- There are no major independent variables we are leaving out
- There are no independent variables included that do not belong
Second: "Linear"
- This is where scatter plots are useful
- Check that each independent variable (predictor) exhibits a linear
relationship with the response
- Straight line
- plot(dependent, independent)
