User: Guest  Login
Document type:
Masterarbeit 
Author(s):
Tom Hochsprung 
Title:
Learning sparse Gaussian graphical models with few covariance queries 
Abstract:
Graphical models are useful in modelling multivariate dependencies; learning their structure based on data is a key problem. Classical methods for learning Gaussian graphical models rely on storing the entire covariance matrix, in high-dimensional settings this is often too expensive. We investigate an algorithm of Lugosi et al. (2021) that is able to learn Gaussian graphical models, in particular partial correlation graphs, in a short time using only a few entries of the covariance matrix. W...    »
 
Subject:
MAT Mathematik 
DDC:
510 Mathematik 
Supervisor:
Mathias Drton 
Advisor:
Carlos Améendola Cerón 
Date of acceptation:
01.10.2021 
Year:
2021 
Quarter:
4. Quartal 
Year / month:
2021-10 
Month:
Oct 
Pages:
204 
Language:
en 
University:
Technische Universität München 
Faculty:
Fakultät für Mathematik 
TUM Institution:
Lehrstuhl für Mathematische Statistik 
Format:
Text