Registration Number
0214U002853,
0111U001105
, R & D reports
Title
Hybrid fuzzy models based identification of multifactor dependences
popup.stage_title
Head
Shtovba Serhiy Dmytrovych,
Registration Date
30-01-2014
Organization
Vinnitsa National Technical University
popup.description2
Object of the research is the multifactor dependences in engineering, economical, medical, and other fields that are available in form of expert-experimental data. Goal of the research is increasing of quality multifactor dependence identification under uncertainty due to a creation of new method for hybrid fuzzy models extraction from expert-experimental data. Methods of the research are systems approach, mathematic modeling, fuzzy sets theory, theory of identification, optimization techniques, and machine learning. The results are: 1) method of assessing and providing the quality of multifactor dependencies identification using fuzzy knowledge base; 2) multivariate dependencies identification method based on joint usage of fuzzy knowledge bases with different formats 3) multivariate dependencies identification method based on joint usage of fuzzy knowledge bases and of fuzzy regression equations. To automate the identification of multivariate dependencies based on hybrid fuzzy models developed a set of basic software modules. Benefits of the proposed methods is demonstrated on real dependencies identification for the following tasks: diagnosis of heart disease; forecasting vehicle fuel efficiency; recognition of Italian wines; prediction of concrete strength; predicting 10-year risk of coronary heart disease, diagnosis of heart diseases; quality assessment of hypertext systems; express detection of stationary sources of excess emissions; forecasting natural gas consumption by individual users; forecasting reliability of computer system man-operator; predicting the cost of software development. Application area is various R&D organizations, industrial enterprises, medical institutions, financial groups, military and governmental structures, which are interested in generalization of their observations under uncertainty and extraction the adequate, compact and understandable mathematical models of complex multi-factor dependences.
Product Description
popup.authors
Бікс Юрій Семенович
Биков Микола Максимович
Катєльніков Денис Іванович
Кисса Олександр Володимирович
Козачко Олексій Миколайович
Мазуренко Віктор Володимирович
Москвін Олексій Михайлович
Нагорна Анастасія Володимирівна
Ракитянська Ганна Борисівна
Савчук Дмитро Анатолійович
Штовба Олена Валеріївна
Штовба Сергій Дмитрович
popup.nrat_date
2020-04-02
