Exercise

# Another bedroom?

In previous exercises, you looked at some features that influence house prices: living area, number of bathrooms, etc. In this exercise, you'll explore the surprising impact of bedrooms and see how to avoid a potentially misleading conclusion. If you want, you can check out another supplemental video here, where I give a real-life example of partial and total change.

Instructions

**100 XP**

- Train a linear architecture model for
`price`

in the`Houses_for_sale`

data, using`land_value`

,`living_area`

,`fireplaces`

,`bathrooms`

and`bedrooms`

as explanatory variables. - Calculate the effect size of living area.
- Calculate the effect size of bathrooms. Use argument
`step = 1`

to increment`bathrooms`

by 1. - Calculate the effect size of bedrooms, also with
`step = 1`

. The result here may strike you as counter intuitive. Think about what it might mean in terms of an actual house to add a bedroom*without*adding living area. - Train another linear model, this time arranging to let
`living_area`

change as it will with the other variables. - For the new model, what's the effect size of bedrooms? Use
`step = 1`

again.