In information storage and retrieval systems, the clustering of input information is necessary. Various definitions and methods of clustering have been put forward. However, here it is defined as the grouping of almost the same kinds in the case of documents, and the method of clustering by multiple correlation coefficients is proposed.