An Extension of the Gap Statistic for Ordinal/Categorical Data
Bhattacharyya distance
Chi-square distance
Discrete application of clusGap
Discrete application of clusGap - core function.
Clustering generating function
Cramer's V modified pairwise vector function based on the function fou...
Cramer's V distance
Bhattacharyya's distance (wrapper)
Chi-square distance (wrapper)
Cramer's V distance (wrapper)
Hamming distance wrapper function
Hellinger distance (wrapper)
Sample-to-sample heatmap
Calculate categorical distance matrix for discrete data
Criteria to determine number of clusters k
Hellinger distance
Adapted k-modes algorithm
Summary Heatmap for categorical data
Discrete Data Heatmap
Simulate Data
The gap statistic approach is extended to estimate the number of clusters for categorical response format data. This approach and accompanying software is designed to be used with the output of any clustering algorithm and with distances specifically designed for categorical (i.e. multiple choice) or ordinal survey response data.