We will split the data to train and test datasets such that the test dataset is $20\%$ of original data. Use sklearn.model$\_$selection.train$\_$test$\_$split function to split the data frame to X$\_$train and X$\_$test. X$\_$train is $80\%$ of observation randomly chosen. X$\_$test is the rest $20\%.$ Both X$\_$train and X$\_$test are pd$.$DataFrame object and include 'price' in the table. Note that the train$\_$test$\_$split can handle data frame as well as array.