Unsupervised identification of nonstationary dynamical systems using a Gaussian mixture model based on EM clustering of SOMs.
Giorgio BiagettiPaolo CrippaAlessandro CurziClaudio TurchettiPublished in: ISCAS (2010)
Keyphrases
- dynamical systems
- non stationary
- unsupervised learning
- self organizing maps
- mixture model
- k means
- clustering algorithm
- differential equations
- expectation maximization
- linear quadratic
- gaussian mixture model
- nonlinear dynamical systems
- adaptive algorithms
- qualitative simulation
- gaussian distribution
- supervised learning
- semi supervised
- state space
- agent environment
- phase space
- maximum likelihood
- gaussian mixture
- autoregressive
- gaussian model
- em algorithm
- linear dynamical systems
- random fields
- generative model
- gaussian kernel
- machine learning