The description of metadata of the multidimensional information systems using test data

This paper examines the possibility of application of test data in the metadata description in information systems constructed on the basis of a multidimensional approach. In case of the description of the characteristics of the observed phenomenon using a large number of aspects, the multidimensional data cube, which is the basis of the information system, is characterised by high sparsity. It complicates the organization of data storage. This paper proposes clustered method of describing the data, which makes it possible to express the semantics of the subject domain. It is necessary to select the groups of members for dimensions that are semantically associated with the groups of members of other dimensions. The relationship between groups of members of different dimensions allows to identify clusters in the data cube, i.e. sets of cells that have similar properties and can be described in the same way. Clusters are used as the main element of information system data model. The problem of metadata description in the information system leads to the problem of setting the parameters of such clusters. Test data can be used in the process of describing the structure of a multidimensional cube. Such structures are data models that express the individual properties of the observed phenomenon. Test data can also be used in the process of testing possible methods of data analysis in multidimensional cube. In the process of development of a multidimensional information system can be used different methods to generate test data to suit the structure of clusters of cells in a multidimensional cube. The first method is applied when setting the values of measures that are semantically not related. The facts in this case are described by the Cartesian product of groups of values of measures. The second method is applied if values of measures correspond to different aspects of the same characteristic. The third method is applied if there is a correspondence between members and values of measures in the facts. © Copyright 2017 for the individual papers by the papers' authors.