magistrsko delo
Lovro Podgoršek (Author), Tomaž Curk (Mentor)

Abstract

Vsako leto se na področju bioinformatike izvede na stotine novih raziskav. Rezultati le teh so razdrobljeni po različnih podatkovnih bazah, ki so med seboj nepovezane, ali pa sploh niso dostopne preko spleta. Vse več znanstvenikov zanima, če bi lahko te podatke združili in izluščili odvisne medsebojne povezave med podatki. V magistrskem delu predlagamo algoritem in podatkovno strukturo za združevanje podatkov ter se osredotočimo na iskanje skritih povezav z večmodalno konvolucijsko nevronsko mrežo tipa samokodirnik. Predlagano rešitev ovrednotimo z algoritmom matričnega razcepa DFMF. V nalogi pokažemo, da stiskanje in razširjanje različnih podatkov v skupen nižje dimenzionalni prostor odkrije odvisne medsebojne povezave med podatki.

Keywords

biološki podatki;zlivanje podatkov;samokodirnik;multimodalna nevronska mreža;matrični razcep;računalništvo;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: [L. Podgoršek]
UDC: 004.8:57(043.2)
COBISS: 1538364867 Link will open in a new window
Views: 635
Downloads: 221
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: Data fusion of biological data using multimodal neural networks and matrix factorization
Secondary abstract: Biological research is conducted yearly in the field of bioinformatics. However, their outcomes and insights remain scattered across different unconnected databases, that are often not accessible online. There is an increased interest in the science community to connect these datasets and uncover potential relationships. The thesis presents an algorithm and data structure for connecting multiple datasets, and thereby focuses on uncovering data relationships with the method of multimodal convolution autoencoder. The solution is evaluated by the DFMF matrix factorization alghorithm. The results show that encoding and decoding data to a common lower dimensional space reveals dependent data relationships.
Secondary keywords: biological data;data fusion;autoencoder;convolutional neural network;matrix factorization;computer science;computer and information science;master's degree;
Type (COBISS): Master's thesis/paper
Study programme: 1000471
Embargo end date (OpenAIRE): 1970-01-01
Thesis comment: Univ. v Ljubljani, Fak. za računalništvo in informatiko
Pages: 72 str.
ID: 11225355