Identifikacija tipova 1D-signala pomoću nenadziranog dubokog učenja

master's thesis

• Title (english): Identification of 1D-Signal Types Using Unsupervised Deep Learning
• Author: Jan Tomljanović
• Mentor: Mile Šikić (mentor)
• Committee member: Mile Šikić (predsjednik povjerenstva)
• Committee member: Marko Subašić (član povjerenstva)
• Committee member: Krešimir Križanović (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2017-07-10, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Tijekom de novo procesa sastavljanja genoma određene tipovi očitanja mogu
uzrokovati probleme prilikom rekonstrukcije genoma. Cilj ovog rada je naučiti više
o mogućim tipovima očitanja te kako klasificirati očitanja koristeći nenadzirano
učenje. Koristeći preklapanja među očitanjima, za svako očitanje generiran je graf
pokrivenosti i oni su dalje analizirani. Autoenkoder je korišten s ciljem sažimanja
signala, tj. grafa pokrivenosti, i zatim je nad sažetim prikazom podataka obavljeno
grupiranje. Korišeni su varijacijski i denoising autoencoder te algoritmi grupiranja
k-means i spektralno grupiranje. Nađene grupe signala su vizualizirane te je
provedena semantička analiza grupa. Procjenjena je kvaliteta klasifikacije signala
korištenjem nenadziranog učenja.

Abstract (english)

During de novo genome assembly process, certain types of sequenced reads can
cause problems during genome reconstruction. Goal of this thesis is to learn more
about possible types of reads and classification of those reads using unsupervised
learning. Coverage graphs of reads are generated using read overlaps and those
coverage graphs are further analysed. Autoencoder is used to compress the
signal, i.e. the coverage graph, and clustering algorithm is then applied to the
compressed data. Variational and denoising autoencoders along with k-means
and spectral clustering algorithms are used. Visualisation of found clusters is
performed along with semantic analysis. Signal classification quality using
unsupervised learning is estimated.

• Keywords: bioinformatika; nenadzirano učenje; duboko učenje; autoenkoderi
• Keywords (english): bioinformatics; unsupervised learning; deep learning; autoencoders
• Language: english
• URN:NBN: urn:nbn:hr:168:474486
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Repository: FER Repository
• Created on: 2019-03-14 01:37:49