Skip to content

Latest commit

 

History

History
17 lines (10 loc) · 539 Bytes

File metadata and controls

17 lines (10 loc) · 539 Bytes

sc-5-nn

The goal is to generate seriation samples from an approximate nearest neighbor model, with highly localized connectivity within a region which is square, but otherwise has the same parameters as sc-4-nn.

Experiment Details

  • 10 slices
  • 32 subpopulations per slice
  • Edges per subpopulation have mean 1.25 and sd 0.25 in a lognormal distribution
  • The regional space is square (aspect ratio 1.0)
  • Edges link vertices chosen randomly but biased by an negative exponential decay kernel with coefficient 3.0