calcSegmentationPower: Calculates the 'Segmentation Power' of the Specified... In bayesMCClust: Mixtures-of-Experts Markov Chain Clustering and Dirichlet Multinomial Clustering

Description

Calculates the 'segmentation power' and optionally the 'sharpness' of the specified classification. The 'segmentation power' corresponds to the maximum individual posterior classification probability. The closer the maximum individual posterior classification probability is to 1, the higher is the segmentation power for individual i. Note that one minus these numbers corresponds to the misclassification risk in each group; hence the closer to one, the smaller is the misclassification risk.

The 'sharpness' on the other hand considers the difference between highest (maximum) and second highest individual posterior classification probabilities, which gives some hints about the 'sharpness' of the classification.

Usage

 ```1 2 3 4``` ```calcSegmentationPower(outList, classProbs, class, printXtable = TRUE, calcSharp = TRUE, printSharpXtable = TRUE, grLabels = paste("Group", 1:outList\$Prior\$H)) ```

Arguments

 `outList` specifies a list containing the outcome (return value) of an MCMC run of `mcClust`, `dmClust`, `mcClustExtended`, `dmClustExtended` or `MNLAuxMix`. `classProbs` A matrix with dimension N x H containing the individual posterior classification probabilities returned by `calcAllocations`. `class` A vector of length N containing the group membership returned by `calcAllocations`. `printXtable` If `TRUE` (default) a LaTeX-style table of the segmentation power is generated/printed. `calcSharp` If `TRUE` (default) also the 'sharpness' is calculated. `printSharpXtable` If `TRUE` (default) the 'sharpness' is also printed (provided that `calcSharp=TRUE`). `grLabels` A character vector giving user-specified names for the clusters/groups.

Details

Reported are summary statistics including the quartiles and the median of the distributions of the segmentation power and the 'sharpness' for all individuals within a certain cluster/group as well as for all individuals.

Value

A list containing:

 `segPowTab ` A matrix containing the segmentation power: reported are summary statistics of the distribution of the maximum individual posterior classification probabilities for all individuals within a certain cluster as well as for all individuals. `sharpTab ` A matrix containing the 'sharpness': reported are summary statistics of the difference between highest and second highest individual posterior classification probabilities within groups and overall. `maxProbs ` A vector containing the maximum individual posterior classification probabilities. `sharp ` A vector containing the differences of the individual maximum and the second highest posterior classification probabilities.

Note

Note, that in contrast to the literature (see References), the numbering (labelling) of the states of the categorical outcome variable (time series) in this package is sometimes 0,...,K (instead of 1,...,K), however, there are K+1 categories (states)!

Author(s)

Christoph Pamminger <[email protected]>

References

Sylvia Fruehwirth-Schnatter, Christoph Pamminger, Andrea Weber and Rudolf Winter-Ebmer, (2011), "Labor market entry and earnings dynamics: Bayesian inference using mixtures-of-experts Markov chain clustering". Journal of Applied Econometrics. DOI: 10.1002/jae.1249 http://onlinelibrary.wiley.com/doi/10.1002/jae.1249/abstract

Christoph Pamminger and Sylvia Fruehwirth-Schnatter, (2010), "Model-based Clustering of Categorical Time Series". Bayesian Analysis, Vol. 5, No. 2, pp. 345-368. DOI: 10.1214/10-BA606 http://ba.stat.cmu.edu/journal/2010/vol05/issue02/pamminger.pdf

`calcAllocations`, `mcClust`, `dmClust`, `mcClustExtended`, `dmClustExtended`, `MNLAuxMix`
 ```1 2``` ```# please run the examples in mcClust, dmClust, mcClustExtended, # dmClustExtended, MNLAuxMix ```