Semantičko indeksiranje i klasifikacija dokumenata

master's thesis

• Author: Filip Janjić
• Mentor: Pavle Goldstein (mentor)
• Committee member: Pavle Goldstein (predsjednik povjerenstva)
• Committee member: Juraj Šiftar (član povjerenstva)
• Committee member: Dijana Ilišević (član povjerenstva)
• Committee member: Igor Pažanin (član povjerenstva)
• Granter: University of Zagreb; Faculty of Science; (Department of Mathematics); Zagreb
• Defense date and country: 2019-02-28, Croatia
• Scientific / art field, discipline and subdiscipline: NATURAL SCIENCES; Mathematics

Abstract

Cilj ovog rada je klasifikacija proteina pomoću algoritma potpornih vektora i n-gramskog rječnika. Prvo su dani osnovni pojmovi iz linearne algebre, euklidskih prostora i teorije optimizacije, potrebni za lakše shvaćanje tema obrađenih u radu. Objašnjena je matematička pozadina načina rada algoritma, s posebnim naglaskom na jezgrene funkcije koje se koriste. Za analizu se koristi deset proteinskih familija, pri čemu je iz svake familije uzeto 300 proteina za trening i 100 proteina za testiranje. Nakon što je napravljen n-gramski rječnik, provedeni su testovi klasifikacije pomoću algoritma potpornih vektora te su uspoređivani rezultati za različite jezgrene funkcije u ovisnosti o parametrima.

Abstract (english)

The aim of this work is to classify proteins using the support vector machine and the n-gram dictionary. First, we present some basic concepts from linear algebra, Euclidean spaces, and optimization theory that are needed later on. A mathematical background of the algorithm has been explained in detail, with a special emphasis on the kernel functions. Ten protein families were used for the analysis, using 300 protein for each family for training and 100 test proteins. After the n-gram vocabulary was made, classification tests were performed using the support vector machine, and comparisons were made for different kernel functions depending on the parameters.

• Keywords: klasifikacija proteina; linearna algebra; euklidski prostori; teorija optimizacije; jezgrene funkcije; n-gramski rječnik
• Keywords (english): protein classification; linear algebra; Euclidean spaces; optimization theory; kernel functions; n-gram dictionary
• Language: croatian
• URN:NBN: urn:nbn:hr:217:302793
• Study programme: Title: Mathematical Statistics; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra matematike (magistar/magistra matematike)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2019-07-19 10:50:49