- nstart > 1 recommended (e.g.
nstart = 25): runs algorithm from multiple random starts
Explore
14,613 skills indexed with the new KISS metadata standard.
nstart = 25): runs algorithm from multiple random starts
not test statistics.
p-values are conservative (too large). Use dgof or ks.test with exact = NULL for discrete distributions.
convergence check fails — use control = list(scaleOffset = 1) or jitter data.
SSasymp
Type I SS (sequential) are computed — order of terms matters.
contr.poly for ordered.
start
isdir
result column is a matrix column inside the data frame.
data
operates on whole rows. For lists
class table. Convert to data frame with as.data.frame(tbl).
not position. Missing columns get NA.
not cbind.matrix. Mixing matrices and data frames can give unexpected results.
names(y))` — can silently merge on unintended columns if data frames share column names.
Date objects become numeric).
apply coerces to matrix via as.matrix first — mixed types become character.
which can silently pick up wrong variables in programmatic use. Use [ with explicit logic in functions.
1]` returns a vector (drop=TRUE default for columns)
j)]`) coerces to matrix first — avoid.
start
skeleton) — flesh` is the flat data
warning = function(w) {...}) catches the first warning and exits. Use withCallingHandlers + invokeRestart(muffleWarning)` to suppress warnings but continue.