Dynamic Mixture Models for Multiple Time-Series
Abstract
Traditional probabilistic mixture models such as Latent Dirichlet Allocation imply that data records (such as documents) are fully exchangeable. However, data are naturally collected along time, thus obey some order in time. In this paper, we present Dynamic Mixture Models (DMMs) for online pattern discovery in multiple time series. DMMs do not have the noticeable drawback of the SVD-based methods for data streams: negative values in hidden variables are often produced even with all non-negative inputs. We apply DMM models to two real-world datasets, and achieve significantly better results with intuitive interpretation.
Cite
Text
Wei et al. "Dynamic Mixture Models for Multiple Time-Series." International Joint Conference on Artificial Intelligence, 2007.Markdown
[Wei et al. "Dynamic Mixture Models for Multiple Time-Series." International Joint Conference on Artificial Intelligence, 2007.](https://mlanthology.org/ijcai/2007/wei2007ijcai-dynamic/)BibTeX
@inproceedings{wei2007ijcai-dynamic,
title = {{Dynamic Mixture Models for Multiple Time-Series}},
author = {Wei, Xing and Sun, Jimeng and Wang, Xuerui},
booktitle = {International Joint Conference on Artificial Intelligence},
year = {2007},
pages = {2909-2914},
url = {https://mlanthology.org/ijcai/2007/wei2007ijcai-dynamic/}
}