Skip to content

Latest commit

 

History

History
25 lines (15 loc) · 599 Bytes

homophily_model.md

File metadata and controls

25 lines (15 loc) · 599 Bytes

Here we present the homophilic scale-free model result

The assumption is that corporate tend to have more link to corporate and

I. node attribute (x,y) where x is U(0,1) and y is [0,0.5] with prob [0.95,0.05]. the homophily score is the euclidian distance

  • power: 3 fit well
  • avg_clustering 0.007
  • corr: -0.003

II, node in a triangle di=1 gao=1, euclidian distance, non-linear preferential 0.8

  • power:3.3 very well fit
  • no clustering 0.0006
  • -0.002

preferential 1.1

  • power 2.9 doen't fit well
  • 0.022

add non linear homophily score 1-sqrt(distance) can increase the homophily