ecms_neu_mini.png

Digital Library

of the European Council for Modelling and Simulation

 

Title:

Multi-Resolution Modelling Of Topic Relationships In Semantic Space

Authors:

Wang Wei, Andrzej Bargiela

Published in:

 

(2009).ECMS 2009 Proceedings edited by J. Otamendi, A. Bargiela, J. L. Montes, L. M. Doncel Pedrera. European Council for Modeling and Simulation. doi:10.7148/2009 

 

ISBN: 978-0-9553018-8-9

 

23rd European Conference on Modelling and Simulation,

Madrid, June 9-12, 2009

Citation format:

Wei, W., & Bargiela, A. (2009). Multi-Resolution Modelling Of Topic Relationships In Semantic Space. ECMS 2009 Proceedings edited by J. Otamendi, A. Bargiela, J. L. Montes, L. M. Doncel Pedrera (pp. 813-819). European Council for Modeling and Simulation. doi:10.7148/2009-0813-0819

DOI:

http://dx.doi.org/10.7148/2009-0813-0819

Abstract:

Recent techniques for document modelling provide means for transforming document representation in high dimensional word space to low dimensional semantic space. The representation with coarse resolution is often regarded as being able to capture intrinsic semantic struc- ture of the original documents. Probabilistic topic mod- els for document modelling attempt to search for richer representations of the structure of linguistic stimuli and as such support the process of human cognition. The topics inferred by the probabilistic topic models (latent topics) are represented as probability distributions over words. Although they are interpretable, the interpreta- tion is not sufficiently straightforward for human under- standing. Also, perhaps more importantly, relationships between the topics are difficult, if not impossible to in- terpret. Instead of directly operating on the latent top- ics, we extract topics with labels from a document col- lection and represent them using fictitious documents. Having trained the probabilistic topic models, we pro- pose a method for deriving relationships (more general or more specific) between the extracted topics in the se- mantic space. To ensure a reasonable accuracy of mod- eling in a given semantic space we have conducted ex- periments with various dimensionality of the semantic space to identify optimal parameter settings in this con- text. Evaluation and comparison show that our method outperforms the existing methods for learning concept or topic relationships using same dataset.

Full text: