Methods for identifying clusters of cells in sparse data cubes of multidimensional information systems

The data model in information system based on the multi-dimensional approach is a multidimensional data cube. Systems with a multi-Aspect description of the subject area are characterized by large data cubes with sparseness. It complicates the data storage organization and creates difficulties in the process of data analysis. Possible cube cells can be represented as possible member combinations. The analysis of semantically related members belonging to different dimensions allows identifying clusters - sets of cells that have similar properties. Clusters are constructed from groups of values of dimensions, which are semantically related to groups of values of other dimensions. Logical methods of intellectual analysis can be used to construct clusters of cells. In the framework of a logical approach, a member combination is represented as a conjunction of the pairs "Measurement" - "Measurement value". The identified clusters can be used as the elements in the data model of the information system. © 2017 CEUR-WS. All rights reserved.