Pages: 187-198
Published: 28.02.2018
Abstract: The article describes the general principles of the current version of the developed virtual database (VDB). The composition of the applied data models (in different versions) was previously used in generating k? and (or) kt pieces of multidimensional (?, Y) -,(C, ?, Y) - (C, ?) -,(?)-samples, m> n (or mt>n) [22-24]. The structures of the input and output r, a unified digital object (UDO) are designed. UDO is the key component of the VDB architecture. The input file of the VDB consists of a small number of UDOs, the output file consists of a very large number of UDOs that are deleted after they are used. No need to store, protect the database. One UDO makes it possible to generate k? or kt pieces of multidimensional data- (?) -samples. The real and model n-dimensional ?-samples have exactly the same values of some parameters from the above-mentioned sets of parameters. The exact equality of the arithmetic mean and variance values for all n variables to given values is a distinctive property of multidimensional samples of the volume m>n - elements of the UDO. Model objects possess many algebraic, geometric, statistical properties of another object (original) and were used instead of it [23-25]. Sufficient (from the values of parameters from three sets [2,8,11]) to the only real standardized multidimensional sample of values of z-variables are modeled. From the multidimensional samples and from the matrices appropriate from the output UDO, the user constructs on the PC (possibly using the values of arithmetic mean and variance for all n variables of his real multidimensional sample) graphical objects: diagrams, drawings, diagrams, digraphs, verbal conclusions (digital knowledge, digital knowledge). This version of WBD is focused on the use of DM PCA, IM PCA in their different versions, different sets of input objects. not only one output object-a multidimensional ?-sample. The files in the database of the usual database are always protected, and in VDB files for storage are not available, but can be generated at any time, depending on the structures of the UDO input file. The validity of the presence of a particular UDO is the key Criterion No. 1 for VDB. The advantage of VDB is the "virtuality" of the database of model ?-samples that are adequate to one or another real multidimensional sample, interpreted as ?-sample (Criterion No. 2).
Key words: virtual database, applied data models, samples.
|
