My research interests have included the fields of artificial intelligence, fuzzy systems, hybrid intelligent systems, and parallel computing. My current interests are in the application of soft computing techniques for modeling and decision making in complex problem domains. I am also interested in automated knowledge extraction from large data sets, machine learning in the presence of uncertain and imprecise data and information retrieval using artificial intelligence techniques.
In the area of fuzzy systems design my efforts during M.S. studies focused on the refinement of fuzzy rule-based models and learning algorithms. Some avenues of research in this direction include rule base reduction to enhance interpretability, adaptation rules with transparency conditions, combining models like neuro-fuzzy, and the analyzing of the approximation properties and interpretability of fuzzy systems. The overall objective is to enhance the robustness and speed of the learning process by the development of hybrid algorithms that employ techniques from several soft computing disciplines, and to develop learning strategy that generates more accurate, concise, and interpretable models with fewer training data volume.
M.Sc. Thesis: “Fuzzy Systems Modeling, Optimization and Automated Learning: Design and Simulation Tools”, June 2004.
Fuzzy set theory, neural networks, and genetic algorithms are three principal constituents of the soft computing methodology that are particularly suited for building and learning system models from training and experiment data. Fuzzy set theory provides a theoretical foundation for modeling complex systems.
The focus of conducted research was the development and analysis of algorithms to generate fuzzy rule bases from training data. In order to fulfill this goal the following tasks were completed: an investigation of the existing approaches to data-driven fuzzy modeling, development of a technique to design fuzzy inference systems of Mamdani type with transparency constraints, and creating a set of software tools for the automated learning and optimization, simulation and verification of fuzzy systems. Developed tools were applied to solve control problems and nonlinear systems identification problems.