Using mcsimex with a misclassification matrix estimated from the fitted clustering model reduces bias in regression coefficients when outcomes are regressed on unsupervised cluster labels.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ML 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Regression on imperfect class labels derived by unsupervised clustering
Using mcsimex with a misclassification matrix estimated from the fitted clustering model reduces bias in regression coefficients when outcomes are regressed on unsupervised cluster labels.