Javascript must be enabled for the correct page display

Non-stationary sparse time series chain graphical models for reconstructing networks

van Ommeren, R.P.W. (2015) Non-stationary sparse time series chain graphical models for reconstructing networks. Bachelor's Thesis, Mathematics.

[img]
Preview
Text
RPW_van_Ommeren_bachelor_thesis.pdf - Published Version

Download (545kB) | Preview
[img] Text
Toestemming.pdf - Other
Restricted to Backend only

Download (54kB)

Abstract

For analyzing high dimensional data, we consider three models. Firstly, the stationary sparse time series chain graphical model is explained. This model is parametrized by a precision matrix and an autoregressive coefficient matrix. We then consider two non-stationary alterations of this model. In the first model, we consider a complete change in the autoregressive matrix at certain change points. In the second model, we consider changes in part of the autoregressive matrix, leaving other parts untouched. Using penalized likelihood with the SCAD penalty, we obtain sparse estimates for the matrices.

Item Type: Thesis (Bachelor's Thesis)
Degree programme: Mathematics
Thesis type: Bachelor's Thesis
Language: English
Date Deposited: 15 Feb 2018 08:04
Last Modified: 15 Feb 2018 08:04
URI: http://fse.studenttheses.ub.rug.nl/id/eprint/12830

Actions (login required)

View Item View Item