###########################################################################################

Multiple linear Regression with Automated Backward Elimination (with p-value and adjusted r-squared)

###########################################################################################

Multiple linear regression model implementation with automated backward elimination (with p-value and adjusted r-squared) in Python and R for showing the relationship among profit and types of expenditures and the states.

Python results:

To make our model reliable and select the features that have an impact on the output, we use Backward elimnation. Since standard R2 is biased,for better performance and model implementation, adjusted r-squared is included in the estimation.

This method allowed to narrow down the dataset features (expenditure types) to 2 and icreasethe regression model score from 0.93 to 0,95.

R results

Implemeting the model in R has shown a bit different results: only one feature (expenditure type) is estimated to be statistically significant.