Abstract: Motivated by genetic expression data, we introduce plaid models. These are a form of two-sided cluster analysis that allows clusters to overlap. Plaid models also incorporate additive two way ANOVA models within the two-sided clusters. Using these models we find interpretable structure in some yeast expression data, as well as in some nutrition data and some foreign exchange data.
Key words and phrases: Microarrays, SVD, transposable data, unsupervised learning.