<div class="csl-bib-body">
<div class="csl-entry">Dittrich, T., & Matz, G. (2022). A Linearly Constrained Power Iteration for Spectral Semi-Supervised Classification on Signed Graphs. In <i>2022 IEEE Data Science and Learning Workshop (DSLW)</i> (pp. 1–6). https://doi.org/10.1109/DSLW53931.2022.9820404</div>
</div>
-
dc.identifier.uri
http://hdl.handle.net/20.500.12708/139239
-
dc.description.abstract
In this work we consider the problem of semi-supervised node classification and extend the method of Xu et al. [1] to a multiclass setting. [1] proposed a linearly constrained variant of the power iteration for classification with two classes where the semi-supervised knowledge was incorporated in the linear constraints. For the multiclass setting we extend this work to a linearly constrained orthogonal iteration. In order to optimize all clusters at the same time, we deviate from the orthogonal iteration by replacing the QR-decomposition for orthonormalization by a projection operation. Our method is parameter-free in the sense that it only relies on a positive definite matrix representation of the graph which admits both conventional graphs and signed graphs. In our experiments we show that the joint optimization using a projection operation outperforms the sequential optimization with orthogonality constraints.
-
dc.language.iso
en
-
dc.subject
linear constraints
en
dc.subject
node classification
en
dc.subject
signed graphs
en
dc.subject
spectral clustering
en
dc.title
A Linearly Constrained Power Iteration for Spectral Semi-Supervised Classification on Signed Graphs
-
dc.type
Inproceedings
en
dc.type
Konferenzbeitrag
de
dc.relation.isbn
978-1-6654-5426-1
-
dc.relation.doi
10.1109/DSLW53931.2022
-
dc.description.startpage
1
-
dc.description.endpage
6
-
dc.type.category
Full-Paper Contribution
-
tuw.booktitle
2022 IEEE Data Science and Learning Workshop (DSLW)
-
tuw.peerreviewed
true
-
tuw.researchTopic.id
I7
-
tuw.researchTopic.name
Telecommunication
-
tuw.researchTopic.value
100
-
tuw.publication.orgunit
E389-03 - Forschungsbereich Signal Processing
-
tuw.publisher.doi
10.1109/DSLW53931.2022.9820404
-
dc.description.numberOfPages
6
-
tuw.author.orcid
0000-0003-1784-806X
-
tuw.event.name
2022 IEEE Data Science and Learning Workshop (DSLW)