Regression Analysis Flashcards
Regression
- One affects the other
- Cause and Effect
- X and Y cannot be changed
- Data represented by line
Correlation
- Relationship
- Variables move together
- X and Y can be interchanged
- Data represented by single point
X Variable
Predictor
- Independent Variable
Y Variable
Response
- Dependent Variable
Slope
- Statistic for every “X” we should see beta change in “Y”
ex: for every MM$ of sales (X) we will see a beta increase in units(Y)
Y-Intercept
-Horsepower vs Gas Mileage
WARNING:
Regression doesn’t PROVE CAUSE and EFFECT, but it can be used to PREDICT cause and effect
SSR = Variation Explained by Regression
for each set of data we take the predicted (y-value - mean of actual y)^2
then we add them all up
SSE = Variation that is Unexplained by Model
for each set of data we
SST = SSR + SSE
total variation around the mean = variation explained by regression + unexplained or error variation
R^2
R^2 is the correlation coefficient
- never have a negative coefficient of determination