Zavgorodnii A. Computer methods for integrated databases soundness and completeness improvement

Українська версія

Thesis for the degree of Candidate of Sciences (CSc)

State registration number

0408U001809

Applicant for

Specialization

  • 05.13.06 - Інформаційні технології

11-04-2008

Specialized Academic Board

Д 64.062.01

National Aerospace University "Kharkiv Aviation Institute"

Essay

The object of research is a process of database integration in automatized control systems. The goal of research is a improvement effectiveness of process of corporative database integration by using redundancy what will allow to improve data soundness and completeness. Primary methods of research are systems analysis, theory of probability methods for soundness and completeness estimation; methods of signal-parametric diagnosis theory, sets theory and functional calculus; algorithms theory and finite state machine theory for record linkage method development; for development knowledge discovery method methods of data analysis, algorithms theory, antinoise coding theory are used. Scientific value of results are determined by developed method of database integration that allow to get acceptable database soundness and completeness with decreasing size of double information input by using structural and information redundancy. Practical value consists in developed information technology that was developed based on created in dissertation methods. For the first time integration method based on heterogeneous database linkage was developed which in contrast to others effectively uses structural, information and temporary redundancy that allow to improve soundness and completeness of information in enterprise control systems. Method of increasing data soundness based on information redundancy is improved by using diagnosis models which allow to take away uncertainty about state of data with the help of indirect signs of errors presence and give ability to improve performance of method. Method of natural information redundancy search is improved by compact view introduction and solve searching in view that give ability to improve performance of method. Has received the further development a method of record linkage on the basis of use string similarity domain-independent model that take into account specific of human data input and possibility of presence abbreviations. Introduction is branch- based. The furtheruse of the results got in dissertation is expedient during database integration when data input software can not be changed or such changes is very difficult.

Files

Similar theses