AMBER — Autoassociative Map Builder for tEmporal Representations

AMBER is an open-source Python library for building, training, and analysing Self-Organizing Maps (SOMs), including recurrent (temporal) SOMs for time-series and biosignal data.

Quick install: pip install amber-som

Getting started

• Getting started
  • Install AMBER
  • Verify the installation
  • Five-line quick start
  • Running the tests

Documentation

• User guide
  • Overview
  • Training a SOM
    • Basic usage
    • Choosing the map size manually
    • Distance metrics
    • Normalisation strategies
    • Weight initialisation
    • Other training options
  • Classification
    • Labelled classification
  • Temporal / Recurrent SOM
  • Feature extraction
    • Available features
  • Iterative map-size selection
  • Visualisation
  • Save and load
  • Application domains and example notebooks
• API reference
  • Map
  • Classification
  • TemporalMap
  • TemporalAnalysis
  • FeatureExtractor
  • IterativeSOM
  • Visualization
  • Distances
    • chebyshev_distance()
    • chebyshev_distance_matrix()
    • correlation_distance()
    • correlation_distance_matrix()
    • cosine_distance()
    • cosine_distance_matrix()
    • cross_correlation_distance()
    • cross_correlation_distance_matrix()
    • dtw_distance()
    • dtw_distance_matrix()
    • euclidean_distance()
    • euclidean_distance_matrix()
    • grid_chebyshev()
    • grid_euclidean()
    • manhattan_distance()
    • manhattan_distance_matrix()

Project

• Changelog

Indices and tables

• Index

• Module Index

• Search Page