Processing math: 100%

A DCLVM with K clusters has edges generated as E[Aijx,θ]θiθjexixj2andxi=2ezi+wi where ek is the kth basis vector of Rd, wiN(0,Id), and {zi}[K]n. The proportionality constant is chosen such that the overall network has expected average degree λ. To calculate the scaling constant, we approximate E[exixj2] for ij by generating random npairs {zi,zj} and average over them.

sample_dclvm(z, lambda, theta, npairs = NULL)

Arguments

z

a vector of cluster labels

lambda

desired average degree of the network

theta

degree parameter

npairs

number of pairs of {zi,zj}

Details

Sample form a degree-corrected latent variable model with Gaussian kernel