TDA

Concepts

Algebraic Topology

TDA

Computational Geometry

Pipeline

  • powerful approach to infer robust qualitative, and sometimes quantitative, information about the structure of data.
  • aims at providing well-founded mathematical, statistical and algorithmic methods to infer, analyze and exploit the complex topological and geometric structures underlying the data
  • Data often represented as point clouds in Euclidean Space or more general metric spaces

Library

GUDHI

  1. Input : finite set of points with a notion of distance
  • Points embedded in metric space
  • Pairwise distance matrix
  • Choice of the metric may be critical to reveal interesting topological and geometric feature of the data.

click to edit

click to edit

  1. Filtration: Simplicial Complex or a nested family of Simplicial Complex that reflects structure of data at different scales
  1. Information Extraction:
  • Method: Persistent Homology
  • Stability to perturbations and noise

Therefore, important to understand the statistical behavior of the inferred features

  1. Use extracted information for inference and further use in Machine Learning

TDA and Statistics