Multiple Imputation in Cluster Analysis
Performs K-means clustering with optional variable selection.
Performs K-means with backward selection.
Performs K-means with forward selection.
Performs K-means with forward selection.
Performs K-means with forward selection.
Performs K-means without variable selection.
Computes euclidean distance.
Computes CritCF.
Creates a midata
object.
Computes initial centroids.
Calculates the ranked selection frequency of the variables.
Computes Manhattan distance.
Computes probabilities of (relabeled) cluster and kappas.
Center data.
Computes centroid.
miclust-package: integrating multiple imputation with cluster analysis
Cluster analysis in multiple imputed data sets with optional variable ...
Shows a graphical representation of the results.
Prints the results.
Prints the summary of results.
Relabel clusters.
Standardize data.
Summarizes the results.
Implementation of a framework for cluster analysis with selection of the final number of clusters and an optional variable selection procedure. The package is designed to integrate the results of multiple imputed datasets while accounting for the uncertainty that the imputations introduce in the final results. In addition, the package can also be used for a cluster analysis of the complete cases of a single dataset. The package also includes specific methods to summarize and plot the results. The methods are described in Basagana et al. (2013) <doi:10.1093/aje/kws289>.