Pčelinji algoritmi za odabir podskupa atributa

master's thesis

• Title (english): Bee algorithms for selecting a subset of attributes
• Author: Lucija Puškarić
• Mentor: Goranka Nogo (mentor)
• Committee member: Goranka Nogo (predsjednik povjerenstva)
• Committee member: Miljenko Marušić (član povjerenstva)
• Committee member: Goran Igaly (član povjerenstva)
• Committee member: Zvonimir Bujanović (član povjerenstva)
• Granter: University of Zagreb; Faculty of Science; (Department of Mathematics); Zagreb
• Defense date and country: 2020-09-30, Croatia
• Scientific / art field, discipline and subdiscipline: NATURAL SCIENCES; Mathematics

Abstract

Glavni cilj ovog diplomskog rada je objasniti kako koristiti meta-heuristiku imena Artificial Bee Colony u rješavanju problema odabira podskupa atributa u disciplini strojnog učenja. Procedura je opisana i primijenjena na neke standardne skupove podataka koji se inače koriste u sličnim istraživanjima. Prva dva poglavlja opisuju teorijsku stranu teme, a druga dva poglavlja predstavljaju implementaciju i rezultate. Na početku prvog poglavlja, opisan je problem odabira podskupa atributa. Nakon toga su pojašnjeni korišteni pojmovi, modeli i koncepti strojnog učenja i takoder standardne metode za rješavanje spomenutog problema. U drugom poglavlju su definirani osnovni pojmovi vezani za meta-heuristike. U nastavku je objašnjen algoritam Artificial Bee Colony. Treće poglavlje opisuje korištene skupove podataka, pripremu podataka, prilagodbu te primjenu pčelinjeg algoritma. Isto tako je navedeno koje druge metode za odabir podskupa atributa su primijenjene na skupove podataka. Rezultati pčelinjeg algoritma, usporedba s rezultatima ostalih metoda i nabrojeni parametri su sadržani u četvrtom poglavlju. Na kraju, zaključeno je da se algoritam Binary Artificial Bee Colony pokazao uspješnim u rješavanju opisanog problema.

Abstract (english)

The main goal of this diploma thesis is to explain how to use a meta-heuristic named Artificial Bee Colony to solve the problem of feature selection in the field of machine learning. The Procedure is described and applied to some standard data sets which are usually used in similar researches. First two chapters describe the theoretical side of the topic, and second two chapters represent implementation and results. At the beginning of the first chapter, the problem of feature selection is described. After that, used terms, models, and conceptions of machine learning are depicted, and also some standard methods for solving the mentioned problem. In the second chapter, basic terms related to meta-heuristics are defined. Artificial Bee Colony is explained afterward. The third chapter presents data sets that are used, data preprocessing, adjustment, and application of the bee algorithm. It is also presented which other methods for feature selection are applied to data sets. Results of the bee algorithm, comparison to results of other methods, and listed parameters are contained in the fourth chapter. In the end, it is concluded that Binary Artificial Bee Colony algorithm proved to be successful in solving the described problem.

• Keywords: meta-heuristika; Artificial Bee Colony; strojno učenje; pčelinji algoritam
• Keywords (english): meta-heuristic; Artificial Bee Colony; machine learning; the bee algorithm
• Language: croatian
• URN:NBN: urn:nbn:hr:217:399784
• Study programme: Title: Computer Science and Mathematics; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra računarstva i matematike (magistar/magistra računarstva i matematike)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2021-03-04 12:38:26