Measuring the performance of the regression model

To measure the performance of a regression model, we can calculate the distance from the predicted output and actual output as a quantifier of model performance. In this calculation, we often use root mean square error (RMSE) and relative square error (RSE) as common measurements. In the following recipe, we illustrate how to compute these measurements from a built regression model.

Getting ready

You need to have completed the previous recipe by fitting the house rental data into a regression model and have the fitted model assigned to the variable lmfit.

How to do it…

Perform the following steps to measure the performance of the regression model:

1. Retrieve predicted values by using the predict function:

   > predicted <- predict(lmfit, data=house)
   

2. Calculate the root mean square error:

   > actual <- house$Sqft 
   > rmse <- (mean((predicted - actual)^2))^0.5
   > rmse
   [1] 66894.34
   

3. Calculate the relative square error:

   > mu <- mean...