Residual plot to validate linear model!


We plot the residuals with the independent variable to check the whether the model is best for our data.
If the data points in the plot are randomly distributed then Linear Model is applicable otherwise not.

Can anybody explain here with proper inference?

Thanks in Advance!



I think you are getting confused in two different aspects. You don’t judge the goodness of a linear model by looking at random distribution of residuals.

There are several metrics to look at goodness of a linear model, R-square and Adjusted R-square in case of regression models.

Residuals are plotted to understand whether the assumptions which have gone in building a linear model hold true or not - so it is validation of assumption rather than goodness of fit.