311 search results - page 18 / 63 » A Constructive-Fuzzy System Modeling for Time Series Forecas... |
CVPR
1999
IEEE
14 years 9 months ago
1999
IEEE
We present a novel mixed-state dynamic Bayesian network (DBN) framework for modeling and classifying timeseries data such as object trajectories. A hidden Markov model (HMM) of di...
CVPR
2007
IEEE
14 years 9 months ago
2007
IEEE
We propose a cord distance in the space of dynamical models that takes into account their dynamics, including transients, output maps and input distributions. In data analysis app...
IWANN
2005
Springer
14 years 1 months ago
2005
Springer
Prediction of time series is an important problem in many areas of science and engineering. Extending the horizon of predictions further to the future is the challenging and diffic...
KDD
2012
ACM
11 years 10 months ago
2012
ACM
Historical prices are important information that can help consumers decide whether the time is right to buy a product. They provide both a context to the users, and facilitate the...
NN
2010
Springer
13 years 6 months ago
2010
Springer
This study shows that a mixture of RNN experts model can acquire the ability to generate sequences that are combination of multiple primitive patterns by means of self-organizing ...