Karteikarten
SP500 Week 6 Choosing Predictors
Karten in diesem Stapel (6)
- 1Why is (multi) collinearity between predictors an issue?One of the predictors or several of the predictors are redundant and are measuring the same thing, and make measuring Betas very difficult.
- 2Low levels of tolerance (less than .1)...Indicate a serious problem, because a predictor doesn't explain enough of the variance, and 90% of the variance is due to the other factors.
- 3What do we do if the levels of collinearity are unnaceptable?- Remove some predictors
- Merge highly correlated predictors (data reduction). - 4What is suppression?Where one variable suppresses the other variables in their irrelevant variability.
They correlate together but create a zero-order correlation with the construct. - 5What are other regression methods?Stepwise methods (Forward and backward)
Block entry (hierarchical regression) - 6—
Informationen
- Beschreibung
- Psychology year 2 Karteikarten am SP500 Week 6 Choosing Predictors , erstellt von chantel Smith am 28/11/2018.
- Erstellt von
- chantel Smith
- Aufrufe
- 0
- Bewertung
Mehr zu Regression Assumptions and Diagnostics