This project applies feed-forward neural networks to the estimation of speciation and extinction rates of select taxa in deep time.
Input data comes from multiple sources and includes both species-specific and time series data. Data specific to this project are: fossil occurrence datasets for reptilia, synapsida, and temnospondyli clades from both the PBDB and from field-work by Princeton Ph.D candidates. Predictors include oceanic data, paleolatitudes, localities, and various other climactic variables.
The PyRate BDNN model is developed by Torsten Hauffe and Daniele Silvestro here
