tclustWithInitialization {optimalFlow} | R Documentation |
A wrapper for the function tclust_H.
tclustWithInitialization( initialization, cytometry, i.sol.type = "points", trimming = 0.05, restr.fact = 1000 )
initialization |
Initial solution for parameters provided by the user. Can be a matrix of data containing observations anc cluster assignations or can be a list spesifying a multivariate mixture of gaussians. |
cytometry |
A matrix or data.frame of dimension n x p, containing the observations (row-wise). |
i.sol.type |
Type of initial solutions in c('points', 'barycenters'). 'points' refers to a classified data matrix, while 'barycenters' to a multivariate mixture. |
trimming |
The proportion of observations to be trimmed. |
restr.fact |
The constant restr.fact >= 1 constrains the allowed differences among group scatters. Larger values imply larger differences of group scatters, a value of 1 specifies the strongest restriction. |
A list with entries:
A numerical vector of size n containing the cluster assignment for each observation. Cluster names are integer numbers from 1 to k, 0 indicates trimmed observations.
Number of clusters actually found.
he value of the objective function of the best (returned) solution.
x <- rbind(matrix(rnorm(100), ncol = 2), matrix(rnorm(100) + 2, ncol = 2), matrix(rnorm(100) + 4, ncol = 2)) ## robust cluster obtention from a sample x asking for 3 clusters, ## trimming level 0.05 and constrain level 12 k <- 3; alpha <- 0.05; restr.fact <- 12 output = tclust_H(x = x, k = k, alpha = alpha, nstart = 50, iter.max = 20, restr = 'eigen', restr.fact = restr.fact, sol_ini_p = FALSE, sol_ini = NA, equal.weights = FALSE, trace = 0, zero.tol = 1e-16) ## cluster assigment output2 <- tclustWithInitialization(data.frame(x, output$cluster), x, 'points', 0.05, 10)