information data processing
A.O. Aseev, A.V. Vassenko
The article offers a variant of improving of data base information processing efficiency by use of an algorithm of initial data preparing, along with a method of reduction of system response to queries. The method includes an extended model of tree-like data representation in a RAM and optimal order of query attributes.
During an analysis of modern approaches to elimination of incompleteness in tuples attributes, a functional dependence based algorithm was chosen. Specifics of data processing during decision making required its improvement both for relational and for symbolic data models. As a result of that a transition mathematical apparatus was designed.
In a view of exponential computational complexity of the algorithm, an analysis of different approaches to optimization of output steps number needed for a query answer was made. The analysis showed that associativity of data representation in a RAM and determination of an optimum order of query attributes enumeration let keep computational cost to an acceptable level.
Results of an experimental check of the method confirmed main theoretical standpoints and operability of the algorithm, model and theorem implementing the methodic. Experimental research showed that method makes initial data preparation 5,5–36 time faster and query answer time 1,6–4,5 times less.