General Additive Mixed Models
One of the projects that our data science-focused interns have been working on on is using a General Additive Mixed Model (GAMM) to investigate the large amount of data we have collected over the past ten years. The benefit of a GAMM is that you can view the effects of each of the different independent variables on the dependent variable. This is incredibly useful for our study which has non-linear attributes; for example, how active bees are over the course of a year, and when considering other variables such as the weather. These variables are normal and wouldn’t indicate a significant decrease in the abundance of wild bees. GAMMs work better than linear or polynomial regression to understand more complex patterns in the data, which are common in ecological studies.
