postPairs: Calculate pairwise co-membership probability

In maoyinan/BayesPC: R package for Bayesian Projection Clustering

Description

Given a vector of cluster numbers and projection clustering output, postPairs calculate the posterior probability of any pair of subjects being clustered in the same group.

Usage

postPairs(
  df_of_draws,
  x_var,
  id_var = "ID",
  dat,
  ls_idxA,
  nIter = 10,
  nDraw = 1000,
  nClusters,
  regQ = 1e-06,
  seed = 1
)

Arguments

df_of_draws	Data frame of simulated LMM output
x_var	Character vector of random effect variables
id_var	Character of id variable
dat	Longitudinal data input
ls_idxA	List of random effect indices to project on
nIter	Number of iterations used in clustering optimization
nDraw	Number of draws to sample from projection clustering output
nClusters	Vector of cluster numbers
regQ	Positive regularization value to add to the diagonal of matrix to be inverted
seed	Random seed to initialize cluster centers

Value

List of 2 items. ls_prob: pairwise probability tables corresponding to random effect projection specified in ls_idxA. Row and column index of table indicate subject ID number in dat. Only the lower triangular matrix is filled. arr_cluster: array of optimized cluster labels based on randomly drawn samples corresponding to chosen random effects.

Examples

data(df_of_draws)
ls_idxA <- list(
  seq(10),
  1:4,
  5:7,
  8:10
)
out_pc <- postPairs(df_of_draws, x_var=paste0("Z", 1:10), id_var="ID", dat=DATASET,
                    ls_idxA, nIter=10, nDraw=2, nClusters=4, regQ=1e-6, seed=1)

maoyinan/BayesPC documentation built on Dec. 21, 2021, 1:48 p.m.