# coefficients from decision regression tree

I have a data on which I apply decision regression tree since it gives better numerical values as predicting house prices. I have almost 45 features in which price is the target value. With linear regression, I get bad guesses while decision regression tree(right one) yields far better guesses as shown in the picture.

So, what I don't understand is how it is helpful. I mean in linear regression I can get a formula after training data as y = intercept+0.343543+0.23423+0.3242+..... But I wonder if I get same/similar representation using decision regression tree? If not, how can it be useful?

I cannot use it simply, that is, I cannot write a function taking 45 parameters. In this function, I can predict the price easily.

I'm confused.

• What's your goal here? Is it to understand how a decision tree works, is it to put the output of your model into production or something else? – Philip Kendall May 29 '19 at 6:12