clusters firms based on their cross-sectional wage distributions

grouping.classify(
  measures,
  ksupp = ceiling((1:(60)^(1/1.3))^1.3),
  nstart = 1000,
  iter.max = 200,
  stop = FALSE,
  verbose = 1,
  cval = 1
)

Arguments

measures	specify the type of measures to use (mean and var, quantiles, etc...)
ksupp	vector of different number of groups to try
nstart	(default:1000) total number of starting values
iter.max	(default:100) max nunmber of step for each repetition
sdata	cross sectional data, needs a column j (firm id) and w (log wage)
Nw	number of points to use for wage distributionsdsd
M	you can pass the matrix measurements, requires also weights W (pass on the truth for instance)