19-09-20 Sizemore, Ann E., 2019

Sizemore, Ann E., Jennifer E. Phillips-Cremins, Robert Ghrist, and Danielle S. Bassett. "The importance of the whole: topological data analysis for the network neuroscientist." Network Neuroscience 3, no. 3 (2019): 656-673.

Original | [Mendeley]

ToC

00. Abstract

01. Intro

01 ¶01

algebraic topology

01 ¶02

neteork topology

Figure 1

fig01

01 ¶03

persistent homology

01 ¶04
01 ¶05

02. When should we use topological data analysis?

02 ¶01
02 ¶02

03. Pieces and parts of the simplicical complex

03 ¶01

simplicial complex

unit of nodes is -simplex

-simplex
nodes

rule:
if is simplex in and , then

-skeleton (): collection of all cimplices w dim at most

-cycle: looped patterns of -simplices

03 ¶02

simplex dist

03 ¶03

face: subset of a simplex
e.g., if is 2-simplex, it contains 1-simplex ,
where is face of

Figure 2

fig02

03.01. Chain groups and boudaries

03.01 ¶01

cavities w/i simplicial complexes as akin to bubbles under water

Box 1. From data to simplicial complex

box1 ¶01

clique complex (flag complex) :
assign -simplex to each -clique

box1 ¶02

nerve complex
mat

box1 ¶03

Vietoris-Rips complex

Figure 3

fig03

03.01 ¶02

binary mat ()

first / zeroth chain groups (, ), 1- / 0-chains

03.01 ¶03

boudary operator

boundary

03.01 ¶04

let (the )

boundary map
binary mat

  • boudary of -chain: sum of vertices
  • boudary of -chain: sum of edges
  • boudary of -chain: shell-forming sum of -simplices

chain group , boudary operator ,

03.01 ¶05

Figure 4

fig04

03.02. Homological algebra

03.02 ¶01

-cycle subspace:

-boudary subspace:
subspace of

03.02 ¶02

04. Homology from complex to complex: persistebt homology

04.01. Filtrations

Figure 5

fig05

04.02. Persistent homology

04.03. Extracting topological feautres

Figure 6

fig06

05. Conclusion