# Writing formulas for GAM models

When using gam() to model outcome as an additive function of the inputs, you can use the s() function inside formulas to designate that you want to use a spline to model the non-linear relationship of a continuous variable to the outcome.

Suppose that you want to predict how much weight (Wtloss) a dieter will lose over a 2-year diet plan as a function of:

• Diet type (categorical)
• Sex (categorical)
• Age at beginning of diet (continuous)
• BMI (body mass index) at beginning of diet (continuous)

You do not want to assume that any of the relationships are linear.

Which is the most appropriate formula?