Napovedovanje Alzheimerjeve bolezni s strojnim učenjem iz kliničnih podatkov

magistrsko delo

Matej Zrimšek (Author), Igor Kononenko (Mentor)

Abstract

V sklopu tekmovanja Alzheimer’s Disease Big Data DREAM Challenge #1 (AD#1) želimo z uporabo odprtega znanstvenega pristopa hitro prepoznati natančne napovedne biomarkerje Alzheimerjeve bolezni, ki jih lahko znanstvene, industrijske in regulativne skupnosti uporabljajo za izboljšanje diagnoze in zdravljenja te bolezni. Z uporabo demografskih, kliničnih in genetskih podatkov ter slikanj z MR, pridobljenih na udeležencih v sklopu pobude Alzheimer’s Disease Neuroimaging Initiative (ADNI), smo ustvarili napovedne modele kognitivnih ocen in napovedali neskladja med kognitivnimi sposobnostmi in amiloidnim bremenom. Z izvlačenjem podatkov smo iz enormne množice podatkov sestavili uporaben nabor le-teh za njihovo uporabo kot učne in testne množice. Razvili smo sistem za hitro in poenoteno obdelavo in optimizacijo pridobljenih podatkov. To nam je koristilo, ker smo za napovedovanje poizkusili najrazličnejše pristope in strategije h grajenju učnega modela ter naredili primerjavo njihove učinkovitosti. Poglobili smo se v mnoge druge raziskave na tem področju in z njimi potegnili smernice ter izvedli primerjavo ugotovitev in rezultatov oziroma učinkovitosti z našo.

Keywords

Alzheimerjeva bolezen;bioinformatika;medicinska informatika;strojno učenje;napovedni model;profiliranje bolezni;napovedovanje bolezni;

Data

Language:	Slovenian
Year of publishing:	2020
Typology:	2.09 - Master's Thesis
Organization:	UL FRI - Faculty of Computer and Information Science
Publisher:	[M. Zrimšek]
UDC:	004.42
COBISS:	45086211
Views:	1319
Downloads:	114
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary title:	Alzheimer's disease prediction with machine learning from clinical data
Secondary abstract:	As part of Alzheimer’s Disease Big Data DREAM Challenge #1 (AD#1), we want to quickly identify accurate predictive biomarkers of Alzheimer’s disease using an open scientific approach that can be used by scientific, industrial and regulatory communities to improve diagnosis and treatment. Using demographic, clinical and genetic data and MR imaging obtained from participants in the Alzheimer’s Disease Neuroimaging Initiative (ADNI), we created predictive models of cognitive assessments and predicted discrepancies between cognitive abilities and amyloid load. With data extraction, we compiled a useful subset of data from an enormous amount of it to use it as a training and test set. We have developed a system for fast and uniform processing and optimization of obtained data. This benefited us because we tried a variety of approaches and strategies to build a learning model for prediction and made a comparison of their effectiveness. We delved into much other research in this field, drew guidelines with them and compared the findings and results or efficiency with ours.
Secondary keywords:	Alzheimerʼs disease;bioinformatics;medical informatics;machine learning;predictive model;disease profiling;disease prediction;
Type (COBISS):	Master's thesis/paper
Study programme:	0
Embargo end date (OpenAIRE):	1970-01-01
Thesis comment:	Univ. v Ljubljani, Fak. za matematiko in fiziko, Oddelek za matematiko, Računalništvo in matematika - 2. stopnja
Pages:	XIX, 112 str.
ID:	12349051