Likelihood is a concept that underlies most statistical modeling that falls under the heading of generalized linear model or GLMs.
When we fit any kind of statistical model to a dataset, the goal is to find solutions that either maximize the likelihood of the data, given the model (under a frequentist, maximum likelihood estimation framework), or maximize the likelihood of the data given the data and some prior information on the value of the parameters (under a more Bayesian framework).

In this tutorial, we will introduce some key concepts and tools for smoothing and visualizing potentially non-linear data. We will focus on local regression techniques for continuous outcomes, e.g. BMI, blood pressure, etc, in in one dimension, e.