Institutional Repository of Vilnius University
Didelio duomenų kiekio klasterizavimas /
Abstract
dc:descriptionData has a special value in today’s world - its analysis allows us to solve various problems in medicine, finance, commerce and other fields. As the amount of generated data increases, an incentive emerges to look for faster ways to learn the structures inherent in the data. The aim of this bachelor’s thesis is to propose efficient algorithms for performing large data clustering tasks taking into account the complexity of the task and the desired clustering accuracy and speed. A comparison of distributed clustering algorithms was performed, PKMeans, MR-BKM, MRGMM and Patchwork were selected for further research. Internal and external evaluation methods that can quickly check the quality of obtained clusters in the case of large amounts of data have been identified. In order to increase the quality of clustering results various data preprocessing techniques were suggested. Finally, an experiment of selected algorithms was carried out, obtained results were evaluated and proposals for clustering large amounts of data were formed.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2020
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Vielavičius, Karolis,
- Contributors dc:contributor
-
- Kybartas, Rimantas
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD81805465&prefLang=en_US