class: center, middle, inverse, title-slide # Lab 05 - Nonlinear Models ### Dr. D’Agostino McGowan --- layout: true <div class="my-footer"> <span> Dr. Lucy D'Agostino McGowan </span> </div> --- ## Recall * Polynomials -- * Step functions -- * Linear splines -- * Cubic Splines -- * Natural Splines ---  --- ## Natural cubic splines _A natural cubic spline extrapolates linearly beyond the boundary knots_ -- * This adds 4 extra constraints and allows us to put **more** internal knots for the same degrees of freedom as a regular cubic spline <img src="img/13/natural.png" height=400></img> --- ## It's really easy to do this in R! -- * You can transform any continuous (numeric) variable into a _natural spline_ by using the `step_ns()` function in a `recipe`. -- * You specify the _degrees of freedom_ that you would like to use for that variable using the `deg_free` parameter -- ```r library(ISLR) auto_rec <- Auto %>% recipe(mpg ~ horsepower) %>% step_ns(horsepower, deg_free = 2) ``` -- .question[ How many degrees of freedom did I allocate to the horsepower variable? ] --- ## It's really easy to do this in R! * You can transform any continuous (numeric) variable into a _natural spline_ by using the `step_ns()` function in a `recipe`. * You specify the _degrees of freedom_ that you would like to use for that variable using the `deg_free` parameter * Instead of specifying `deg_free` you could use `tune()` to tune this. -- ```r auto_rec <- Auto %>% recipe(mpg ~ horsepower) %>% step_ns(horsepower, deg_free = tune()) ``` --- class: inverse ## <i class="fas fa-laptop"></i> `Lab 05 - Non-linear Models` * Find Lab 05 on our website * Pull your starter files from GitHub into RStudio * **Note the due date** I have intentionally made this lab short so that it can be completed by Thursday