The University of Cape Town
Honours Minor Dissertation
Gaussian Processes for Time Series Modelling
Raphaela Azah and Sbonelo Gumede
Professor Birgit Erni
There are many ways to model the autocorrelation structure in time series data. Capturing the correlation structure through a Gaussian Process is a non-parametric approach which is very flexible and a good approach for capturing uncertainties. For this project we use Gaussian processes for modelling time series data, simulate from Gaussian processes, reproduce existing model fits to data, and then model a local data set using Gaussian processes.
We apply GPs on air pollution forecasting and gold price modelling.
honours-project/
├── air-quality-forecasting/ # Air pollution forecasting using GP models
│ ├── README.md # Project-specific documentation
│ └── ... # Project files
│
├── gold/ # Gold price modelling using GP models
│ ├── README.md # Project-specific documentation
│ └── ... # Project files
│
├── .gitignore # Git ignore rules for repository
└── README.md # This file
Each project directory contains its own README with specific instructions for:
- Data requirements
- Package installation
- Running the analysis
- Interpreting results
Navigate to the respective project directory for detailed documentation:
- R >= 4.0.0
- Required packages are listed in each project's README
- HPC access recommended for computationally intensive GP models
University of Cape Town - Academic Use