In: Statistics and Probability
What is pitfall? What is the requirement in a regresstion model?
Pitfall : (Pitfall means hidden difficulty which is not expected and also unsusceptible.
In regression model the word pitfall used as regression pitfall that defines the difficulty in the specified regression model.
Some well known regression pitfall are
a. Non constant variance
b. Auto correlation in the model
c. overfitting
d. multicollinearity problem etc
In linear regression we take some assumptions like –
Model is linear in parameters.
Homoscedasticity or equal variance of residuals .
Residuals are uncorrelated