Goodness-of-fit test for latent block models

Latent block models are used for probabilistic biclustering, which is shown to be an effective method analyzing various relational data sets. However, there has been no statistical test determining the row and column cluster numbers of latent models. Recent studies have constructed statistical-test-based methods stochastic models, assume that observed matrix a square symmetric assignments same rows columns. In this study, we developed new goodness-of-fit whether fits given set numbers, or it consists more clusters in at least one direction column. To construct method, result from random theory sample covariance matrix. We experimentally demonstrated effectiveness proposed by showing asymptotic behavior statistic measuring accuracy.