Triclusters of Close Values for the Analysis of 3D Data
Abstract: The paper deals with the problem of triclustering in multivalued triadic contexts in termsof one multidimensional extension of formal concept analysis; triclustering can be viewed as asearch for dense subtensors in three-dimensional tensors over the field of real numbers. Twomethods are proposed for solving this problem, namely, NOAC—a version of the OACtriclustering method for numerical data based on delta operators—and a triadic version ofthek-means method with an improved metric based on Manhattan distance andproximity predicates in each of the three dimensions. Numerical experiments are carried out bothon real and synthetic data and confirm the superiority of the NOAC method in terms of theperformance criteria for the resulting triclusters. © 2022, Pleiades Publishing, Ltd.