Javascript must be enabled for the correct page display

Time series classification using Hankel matrix based dissimilarity measures in Learning Vector Quantization

Holdijk, Lars (2018) Time series classification using Hankel matrix based dissimilarity measures in Learning Vector Quantization. Bachelor's Thesis, Computing Science.

[img]
Preview
Text
main.pdf

Download (1MB) | Preview
[img] Text
toestemming.pdf
Restricted to Registered users only

Download (78kB)

Abstract

Time-series classification is an interesting and challenging sub-domain of classification problems. In distance based classification algorithms, the information hidden in the ordering of the time-series and the possibility of misalignment require the use of specialized dissimilarity measures. In this thesis we look at three such measures, all of which are based on Hankel matrices and the assumption that Hankel matrices with the same subspace originate from the same LTI-series and consequently from the same class. In previous work all three dissimilarity measures have shown competitive results when combined with k-Nearest Neighbours. In our work we combine two of the three dissimilarity measure with Generalized Learning Vector Quantization (GLVQ) using a rewriting of derivatives presented in earlier work. The results presented show promise for the dissimilarity measures to be applied in GLVQ.

Item Type: Thesis (Bachelor's Thesis)
Supervisor:
Supervisor nameSupervisor E mail
Biehl, M.M.Biehl@rug.nl
Degree programme: Computing Science
Thesis type: Bachelor's Thesis
Language: English
Date Deposited: 23 Jul 2018
Last Modified: 20 Aug 2018 12:47
URI: http://fse.studenttheses.ub.rug.nl/id/eprint/18023

Actions (login required)

View Item View Item