| Authors |
|
| Publication date |
2009
|
| Host editors |
-
M. van Erp
-
H. Stehouwer
-
M. van Zaanen
|
| Book title |
Benelearn 09: the 18th Annual Belgian-Dutch Conference on Machine Learning: proceedings of the conference
|
| Event |
Belgian-Dutch Conference on Machine Learning (Benelearn 09), Tilburg, the Netherlands
|
| Pages (from-to) |
7-12
|
| Publisher |
Tilburg: Tilburg centre for Creative Computing (TiCC), Tilburg University
|
| Organisations |
-
Faculty of Science (FNWI) - Informatics Institute (IVI)
|
| Abstract |
In this paper we show how to build a model of ship trajectories in a certain maritime region and use this model to predict future ship movements. The presented method is unsupervised and based on existing compression (line-simplification) and clustering techniques. We evaluate the model with a simple prediction task. In this task we compare the performance of our method to two baseline predictors and show that it outperforms them on average. However, it also shows room for improvement, for which we give a number of suggestions.
|
| Document type |
Conference contribution
|
| Language |
English
|
| Published at |
http://benelearn09.uvt.nl/Proceedings_Benelearn_09.pdf
|
|
Permalink to this page
|