This is a project that investigates the mass loss of red giant stars in the Messier 4 globular cluster. This is done through the use of a neural network acting as a GARSTEC emulator (essentially a rapid predictor of stellar evolution), embedded within a Hierarchical Bayesian Model (HBM).
A repository that documents the final year project work of Kienan Ruddock and Tom Green at the University of Birmingham.
This repository is split into 3 branches:
- Master: this contains nothing
- Kienan: this contains Kienan's work. This is mostly focused on the neural network aspect of the project.
- Tom: this contains Tom's work. This is mostly focused on the HBM aspect of the project.
Note: there is some overlap between work on eachother's branches, this is because of the nature of the project. The HBM requires the neural network to function.
To view work related to the neural network navigate to -> Kienan branch -> Neural Network -> GARSTEC Emulator - GarstecModel_VX (higher number is more recent)
To view work related to the HBM navigate to -> Tom branch -> HBM stuff: this contains a mixture of work using MCMC samplers within both hierarchical and non-hierarchical Bayesian models