The proposed project deals with the development of a new algorithm for adaptive clustering. Research’s motivation starts from the necessity of systems classification during development and planning of development. The new proposed algorithm will be able to adapt to changing system
features and available system characteristics.
The research topic aims to find the techniques to analyse similarities in systems‘ descriptions and investigate if algorithms based on similar projects improve estimation ability. The topic focuses on the analysis and design of data segmentation methods and clustering approaches, then evaluates kernel functions for penalising data items in datasets (windows, weighted windows approach).
The goal of the topics is to create and implement mathematical models for the selection of feature subset from system characteristics.