My research group develops data mining, visualization, and statistical packages. We particularly
focus on developing R packages.
bclust: Bayesian High-Dimensional Clustering using Spike-and-Slab model. Builds a dendrogram using log posterior as a natural distance defined by the model and meanwhile weights the clustering variables. It is also capable to computing equivalent Bayesian discrimination probabilities. The adopted method suites small sample large dimension setting.
labeltodendro: Draws a clustering tree out of labels and heights.
lbiassurv: length-bias correction to the survival function estimation. In a sample, the individuals with short survival die earlier and appear fewer in the sample. This natural phenomenon induces bias in the survival curve estimation. This package offers various corrections to such bias.
baybi: Bayesian spike-and-slab biclustering for repeated measurements.
mixvarselect: Variable selection in mixture models using penalization.
hbiclust: Fast Bayesian hierarchical biclustering and forestogram.