Please use this identifier to cite or link to this item: https://scholarbank.nus.edu.sg/handle/10635/150306
Title: ROBUST EUCLIDEAN DISTANCE MATRIX MODELS FOR EUCLIDEAN EMBEDDING PROBLEMS WITH CORRUPTED DATA
Authors: CHEONG YU JIA
Keywords: Optimization, Euclidean Distance Matrix, Alternating direction method of multipliers, Molecular Conformation Problem
Issue Date: 24-Aug-2018
Citation: CHEONG YU JIA (2018-08-24). ROBUST EUCLIDEAN DISTANCE MATRIX MODELS FOR EUCLIDEAN EMBEDDING PROBLEMS WITH CORRUPTED DATA. ScholarBank@NUS Repository.
Abstract: The molecular conformation problem consists of reconstituting the 3D configuration of proteins from inter-atom distances. This is an application of the Euclidean embedding problem, which is often formulated using Euclidean distance matrix models, and is a challenging problem to solve when the data is sparse, noisy and corrupted, as is the case with nuclear magnetic resonance spectroscopy data. In this masters project, we first formulate convex relaxations of energy based optimization models that make use of the Euclidean distance matrix representation using the alternating direction method of multipliers. Then, we adapt the Penalized Robust Euclidean Embedding model, whose algorithm consists in solving a majorized subproblem that has a closed form solution at each iteration, to handle corrupted data. Both models use a heuristic to detect possible corruptions, thus eliminating the need for manual cleaning of data. These models are then evaluated on sparse, noisy, and corrupted protein configuration data.
URI: http://scholarbank.nus.edu.sg/handle/10635/150306
Appears in Collections:Master's Theses (Open)

Show full item record
Files in This Item:
File Description SizeFormatAccess SettingsVersion 
CheongYJ.pdf1.52 MBAdobe PDF

OPEN

NoneView/Download

Page view(s)

82
checked on Jul 10, 2020

Download(s)

22
checked on Jul 10, 2020

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.