Attestation committee
Accreditation committee
Expert committee
Dispositions, instructions
Normative acts
Nomenclature
Institutions
Scientific councils
Seminars
Theses
Scientific advisers
Scientists
Doctoral students
Postdoctoral students
CNAA logo

 română | русский | english


Decision support on small size passive samples


Author: Popukaylo Vladimir
Degree:doctor of informatics
Speciality: 01.05.04 - Mathematical modelling, mathematical methods, software
Year:2017
Scientific adviser: Svetlana Cojocaru
doctor habilitat, professor, Institute of Mathematics and Computer Science of the ASM
Institution: Institute of Mathematics and Computer Science of the ASM

Status

The thesis was presented on the 22 September, 2017 at the meeting of the Scientific Council and now it is under consideration of the National Council.

Abstract

Adobe PDF document0.46 Mb / in romanian
Adobe PDF document0.49 Mb / in russian

Thesis

CZU

Adobe PDF document 3.09 Mb / in russian
168 pages


Keywords

mathematical modeling, small sample, passive experiment, correlation analysis, regression analysis, outliers, the method of point distributions

Summary

Structure of the thesis: Introduction, three Chapters, General conclusions and recommendations, bibliography of 160 titles, 4 appendices, 120 pages of body text, 30 figures, 23 tables. According to the dissertation research materials 16 printing works are published.

Field of study: small size passive samples obtained in different conditions.

Research objective: to develop a construction technique of adequate mathematical models for small size passive samples, in conditions when classical probabilistic-statistical methods do not allow obtaining valid conclusions.

Research tasks: 1) to analyze various approaches and methods of small size samples processing; 2) to investigate the possibility of using existing criteria for outliers finding in small size samples; 3) to investigate the possibility of the linear correlation detection in the small size samples; 4) to develop a methodology for constructing decision support systems in cases of passive experiment and impossibility of obtaining a large amount of raw data; 5) to test the developed technique on the obtained data in various conditions; 6) to substantiate and to prove the applicability of the developed technique.

The scientific novelty and originality of the research: the original methodology for construction the decision support systems which give the most accurate recommendations for decision-makers basing on passive small samples.

The relevance of research: there are many areas where it is impossible to obtain large amounts of data, and therefore, the decision-making using classical techniques is extremely difficult or impossible.

Important scientific problem solved in the work: the determination of probabilistic and statistical methods of decision support based on small size samples obtained during the passive experiment.

The theoretical significance of the study lies in the methodology of improving the quality of decisions taken on the basis of the proposed approaches and algorithms based on probabilistic and statistical methods of data processing

The practical significance of the research results. The decision support methods, proposed in the thesis, can be widely used in the statistical analysis in various fields of research, where it is impossible to obtain a large amount of data.

The research results implementation: The results of the research is implemented in the direct care of Department of Endoscopic and Minimally Invasive Surgery of the Republican Clinical Hospital (Tiraspol), into educational process of the department "Biology and Human Physiology" and "Therapy №2» Medical Faculty of the T.G. Shevchenko University, into the work of the Research Laboratory "Mathematical Modeling" and academic Master's program in "Computer Science and Engineering" of the same university.