Ugotavljanje genskih prediktorjev s pomočjo inteligentnih sistemov

(magistrsko delo)

Nejc Rednjak (Author), Peter Kokol (Mentor)

Abstract

Pri strojnem učenju (angl. Machine Learning) gre za pridobivanje znanja na podlagi izkušenj. Ne gre za učenje na pamet, ampak za iskanje pravil v učnih podatkih. Najbolj znane metode strojnega učenja so odločitvena drevesa (DT), metoda podpornih vektorjev (SVM) in nevronske mreže (NN). Metode strojnega učenja so nam v pomoč pri ugotavljanju genskih prediktorjev. Algoritmi strojnega učenja imajo prav tako pomembno vlogo pri diagnosticiranju rakavih obolenj. V magistrskem delu smo opisali najbolj znane metode strojnega učenja in jih preizkusili na podatkovni bazi AP_Colon_Kidney. Uporabili smo podatkovno bazo iz spletne zbirke GEMLeR, ki vsebuje podatke o genski ekspresiji za več kot 2000 vzorcev tumorjev. Raziskali smo tudi, kateri geni so najbolj izraženi v primeru rakavega obolenja debelega črevesa in ledvic.

Keywords

strojno učenje;odločitvena drevesa;nevronska mreža;metoda podpornih vektorjev;gen;podatkovna baza;

Data

Language:	Slovenian
Year of publishing:	2015
Typology:	2.09 - Master's Thesis
Organization:	UM FERI - Faculty of Electrical Engineering and Computer Science
Publisher:	[N. Rednjak]
UDC:	575(043.2)
COBISS:	2144932
Views:	1733
Downloads:	137
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary title:	DETERMINING GENETIC PREDICTORS BY USING INTELLIGENT SYSTEMS
Secondary abstract:	With machine learning we acquire knowledge based on experience. It is not about learning by memorization but to search the rules in the learning data. The most known representatives of machine learning are decision trees (DT), support vector machines (SVM) and neural networks (NN). Machine learning methods help us to identify genetic predictors. The machine learning algorithms also play an important role in cancer diagnosis. In our master thesis we describe the most known machine learning methods and test them on an AP_Colon_Kidney database. For these master thesis we have used a database from an GEMLeR online collection which contains data on gene expression with more than 2,000 samples of tumors. We have also investigated which genes are the most xpressed in the case of colon and kidney cancer.
Secondary keywords:	machine learning;decision trees;neural networks;support vector machine;gene;database;
URN:	URN:SI:UM:
Type (COBISS):	Master's thesis/paper
Thesis comment:	Univ. v Mariboru, Fak. za zdravstvene vede
Pages:	IX, 45 f.
ID:	8751540