M. Lukauskas “Development of robust clustering algorithms” doctoral dissertation defense

Author, Institution: Mantas Lukauskas, Kaunas University of Technology

Science area, field of science: Natural Sciences, Informatics, N009

Scientific Supervisor: Prof. Dr. Tomas Ruzgas (Kaunas University of Technology, Natural Sciences, Informatics, N009)

Dissertation Defense Board of Informatics Science Field:
Prof. Dr. Hab. Rimantas Barauskas (Kaunas University of Technology, Natural Sciences, Informatics, N009) – chairperson
Senior Researcher Dr. Gražina Korvel (Vilnius University, Natural Sciences, Informatics, N009)
Dr. Mantas Mikaitis (University of Leeds, United Kingdom, Natural Sciences, Mathematics, N001)
Prof. Dr. Alfonsas Misevičius (Kaunas University of Technology, Natural Sciences, Informatics, N009)
Prof. Dr. Agnė Paulauskaitė-Tarasevičienė (Kaunas University of Technology, Natural Sciences, Informatics, N009)

Dissertation defense meeting will be at Rectorate Hall of Kaunas University of Technology (K. Donelaičio 73-402, Kaunas)

The doctoral dissertation is available at the library of Kaunas University of Technology (Gedimino 50, Kaunas) and on the internet: M. Lukauskas el. dissertation (PDF)

Annotation: This dissertation examines the development and evaluation of robust and efficient data clustering methods. Given the changing importance and need of data analysis, clustering algorithms are becoming an essential tool for solving complex data processing and interpretation problems in various fields such as bioinformatics, image processing, natural language processing, and social network analysis. The aim of the thesis is to develop and investigate robust data clustering methods that can efficiently and accurately group heterogeneous data even in the presence of noise or extreme values. The study analyzed currently used clustering methods, their working principles and common problems such as noise and missing data. To solve these problems, new clustering algorithms and their modifications have been developed, which have proven to be superior to traditional methods. Algorithms proposed in the dissertation are based on the density estimate of the inversion formula, which show higher accuracy compared to existing methods. The performed comparative analysis reveals that the new methods, including their various modifications, achieve better results compared to traditional methods such as k-means clustering. Studies have shown that these methods not only group data more accurately, but also have greater resistance to noise and extreme values. The methods described in the dissertation were applied to solving real problems, including the analysis of economic data, the analysis of text job advertisements and their clustering. The new methods have proven extremely useful for large unstructured data sets, although in some cases their application is limited by high computational costs. The results of the dissertation research were presented at international scientific conferences and published in scientific articles, and new algorithms found practical application in business and scientific projects. In addition, the presented recommendations for the application of the methods were cited and used in the works of other researchers.