Code-Base for the Application of Pyro’s Stochastic Variational Inference on Bayesian Spectral Density Estimation

Prerequisites

To generate the underlying data, R is required, with the packages found in requirements_R.R. Simply run the script in R to install them.

Rscript requirements_R.R

To run the stochastic variational inference in Python, ensure that all required packages are installed, e.g. using

pip install -r requirements_python.txt

Running the Inference Process

First, ensure that the required data is present, by running the R script generate_data.R

Rscript generate_data.R

Then, run the python script main.py.

python main.py

To change the dataset that is used during inference, modify the variable dataset; valid options are sunspots and ma_simulation.