After the optimum dimensionality has been determined in the analysis phase, the RME can compare all of the pages and split them into clusters. The number of clusters is discovered automatically by the RME and is designed to represent the number of natural categories that occur in the data set.
After the clusters are created, the RME creates a title for each one by finding the most significant keywords in each cluster's documents. Clicking on the title of a cluster will automatically generate a data report for that cluster.