lsqfitgp

This is the documentation for the python module lsqfitgp. It provides a simple but powerful interface to use Gaussian processes for fitting. The code is open source, available on github under the GNU General Public License.

Briefly, Gaussian processes are a statistical tool for fitting unknown functions. If you want to know more about Gaussian processes, the book Gaussian Processes for Machine Learning is a good reference available for free. However, understanding the technical details is not necessary to use the basic functionality of the module.

To start, read the Installation section, and then First example: a sine. To report bugs or request features, open an issue on github.

Contents:

• User guide
  • 1. Installation
  • 2. First example: a sine
  • 3. More on kernels
  • 4. Taking derivatives
  • 5. Taking integrals
  • 6. A custom kernel: text classification
  • 7. Multidimensional input
  • 8. Partial derivatives
  • 9. Multidimensional output
  • 10. Splitting components
  • 11. Hyperparameters
  • 12. Hyperparameters in the input
  • 13. Nonlinear models
  • 14. Optimization
• Reference manual
  • 1. Kernels reference
  • 2. Generic kernel classes
  • 3. Kernel decorators
  • 4. Kernel operations
  • 5. Derivative specification
  • 6. Gaussian process class
  • 7. Structured arrays wrapping
  • 8. Fitting
  • 9. Random sampling
  • 10. Copula factories
  • 11. BART
• Examples
  • 1. Example scripts

Indices and tables

• Index

• Search Page