Time Series Relevance Determination Through a Topology-Constrained Hidden Markov Model
Most of the existing research on multivariate time series concerns supervised forecasting problems. In comparison, little research has been devoted to unsupervised methods for the visual exploration of this type of data. The interpretability of time series clustering results may be difficult, even in exploratory visualization, for high dimensional datasets. In this paper, we define and test an unsupervised time series relevance determination method for Generative Topographic Mapping Through Time, a topology-constrained Hidden Markov Model that performs simultaneous time series data clustering and visualization. This relevance determination method can be used as a basis for time series selection, and should ease the interpretation of the time series clustering results.
KeywordsTime Series Feature Selection Latent Variable Model Multivariate Time Series Unsupervised Method
Unable to display preview. Download preview PDF.
- 6.Bishop, C., Hinton, G., Strachan, I.: GTM through time. In: IEEE Fifth Int. Conf. on Artif. Neural Net., Cambridge, U.K., pp. 111–116 (1997)Google Scholar
- 7.Olier, I., Vellido, A.: Capturing the dynamics of multivariate time series through visualization using Generative Topographic Mapping Through Time. In: IEEE ICEIS 2006, Islamabad, Pakistan (2006)Google Scholar
- 9.Vellido, A., Lisboa, P.J.G., Vicente, D.: Robust analysis of MRS brain tumour data using t-GTM. Neurocomputing Accepted for publication (in press)Google Scholar