Welcome to spectrally-regularised-LVM’s documentation!

spectrally-regularised-LVMs is a Python library for linear latent variable models and uses a spectral constraint to ensure that the extracted sources are unique. spectrally-regularised-LVMs has its documentation hosted on Read the Docs.

Overview

Check out the Installation section for any information on how to install the project and the Getting started section for information on how to get the package up and running. The Package guide section contains information related to the main modules in the package, with Modules containing a detailed description of each of the modules with their associated descriptions for all classes, methods, and functions. The Tutorials section contains a more detailed tutorial breakdown for the main aspects of this package.

Note

This project is under active development.

Contents

• 1. Installation
  • 1.1. Pip installation
  • 1.2. Cloning from Github
• 2. Getting started
  • 2.1. Time-series data hankelisation
  • 2.2. Defining an objective function
  • 2.3. Initialising the LVM
  • 2.4. Estimating the LVM parameters
  • 2.5. Using the LVM
• 3. Tutorials
  • 3.1. Tutorial 1: Using a LVM
  • 3.2. Tutorial 2: Defining your own objective function
  • 3.3. Tutorial 3: Using the default objective functions
• 4. Package guide
  • 4.1. negen_approx
  • 4.2. cost_functions
  • 4.3. spectral_regulariser
  • 4.4. helper_methods
  • 4.5. spectrally_regularised_model
  • 4.6. References
• 5. Modules
  • 5.1. Negentropy approximators
  • 5.2. Cost functions
  • 5.3. Spectral regulariser
  • 5.4. Helper methods
  • 5.5. Spectrally regularised model
• 6. Parameter estimation
  • 6.1. Part one: The spectral regularisation term
  • 6.2. Part two: Optimisation formulation
  • 6.3. Part three: Putting it all together

Indices and tables

• Index

• Module Index