makeConstructMatrix: Row-normalizes a matrix with arbitrary weighting for the...

In jcfisher/latentnetDiffusion: Runs latent space diffusion model (ALPHA TESTING VERSION)

Description

Row-normalizes a matrix with arbitrary weighting for the diagonal

Usage

makeConstructMatrix(net, a)

Arguments

net:	a statnet network object or a square matrix
a:	a numeric value indicating how much weight to give to self vs. other. 1 indicates full self-weight (diagonal), 0 indicates full other- weight (off diagonal). Can be a vector to give each person different weights.

Value

a square matrix whose rows sum to 1 or 0

Note

Calculates the construct matrix from Friedkin and Johnson (2011:40). That matrix is defined as AC + I - A, where:

C_{ij} = \frac{W_{ij}}{∑_{k: i \ne k} W_{ik}}

C_{ii} = 0, A_{ii} = a_i, and A_{ij} = 0 for all i \ne j.

Among indegree isolates, ∑_{k: i \ne k} W_{ik} = 0. In those cases, C_{ij} \coloneqq 0, and a_{i} \coloneqq 1.

References

Friedkin, Noah E. and Eugene c. Johnson. 2011. Social Influence Network Theory: A Sociological Examination of Small Group Dynamics. Cambridge University Press: New York.

jcfisher/latentnetDiffusion documentation built on May 20, 2019, 5:26 p.m.