magistrsko delo
Kristjan Reba (Author), Matej Guid (Mentor), Janez Konc (Co-mentor)

Abstract

Iskanje maksimalne klike spada med dobro raziskane NP-polne probleme. Za praktično uporabnost algoritmov za iskanje maksimalne klike morajo biti ti dovolj hitri na ciljni domeni grafov. V zadnjih letih je bilo narejenega veliko napredka na področju strojnega učenja na grafih. V magistrskem delu uporabimo moderne pristope strojnega učenja na grafih za pohitritev dinamičnega algoritma za iskanje maksimalne klike. Pohitritve testiramo na različnih vrstah grafov s poudarkom na različnih vrstah proteinskih grafov. Ugotovimo, da so pohitritve možne in jih lahko dosežemo z dobro izbiro modela za strojno učenje. Ugotovimo tudi, da pohitritve niso velike, vendar pa so konsistentne na skoraj vseh predstavljenih grafih.

Keywords

proteinski graf;maksimalna klika;računalništvo in informatika;magisteriji;

Data

Language: Slovenian
Year of publishing:
Typology: 2.09 - Master's Thesis
Organization: UL FRI - Faculty of Computer and Information Science
Publisher: [K. Reba]
UDC: 004.85(043.2)
COBISS: 82279427 Link will open in a new window
Views: 187
Downloads: 30
Average score: 0 (0 votes)
Metadata: JSON JSON-RDF JSON-LD TURTLE N-TRIPLES XML RDFA MICRODATA DC-XML DC-RDF RDF

Other data

Secondary language: English
Secondary title: Improvements to the dynamic algorithm for finding maximum clique in a protein graph using machine learning
Secondary abstract: Finding maximum clique is a well-researched NP-complete problem. For the practical applicability of algorithms for finding the maximum clique, they must be fast enough on the target domain of graphs. There has been a lot of progress made in recent years in the field of machine learning on graphs. In the master's thesis we use modern approaches to machine learning on graphs to speed up the dynamic algorithm for finding the maximum clique. Speedups are tested with different types of graphs with an emphasis on different types of protein graphs. We find that speeding up the maximum clique search is possible and can be achieved with a good choice of machine learning model. We also find that the speedups are not large but are consistent on almost all the graphs presented.
Secondary keywords: protein graph;maximum clique;machine learning;computer science;computer and information science;master's degree;Strojno učenje;Računalništvo;Univerzitetna in visokošolska dela;
Type (COBISS): Master's thesis/paper
Study programme: 1000471
Thesis comment: Univ. v Ljubljani, Fak. za računalništvo in informatiko
Pages: 66 str.
ID: 13700563