Predicting the Distribution of U.S. Residential Solar Arrays in the U.S. using Zero-Inflated Regression Models by Gracie Guidotti ’23, Statistics Colloquium, Wednesday, February 8, 1:10 – 1:50 pm, North Science Building 114, Wachenheim.
Abstract: How can we model count data with a high amount of zeros? We will explore how the zero-inflated negative binomial regression model can be used to model over-dispersed and zero-inflated count data. We will discuss the theory behind this model and explore how the expectation-maximization (EM) algorithm can be used to estimate the model parameters. Lastly, we will examine how this model can be applied to a zero-inflated dataset of residential solar PV installations in the US.