Estimating state-space models in innovations form using the expectation maximisation algorithm.
Adrian WillsThomas B. SchönBrett NinnessPublished in: CDC (2010)
Keyphrases
- probabilistic model
- dynamic programming
- state space
- learning algorithm
- search space
- computational complexity
- detection algorithm
- times faster
- cost function
- worst case
- improved algorithm
- computationally efficient
- markov chain
- high accuracy
- experimental evaluation
- state variables
- parameter estimation
- expectation maximisation
- k means
- optimal solution
- clustering method
- data sets
- parametric models
- neural network
- similarity measure
- bayesian framework
- objective function
- matching algorithm
- ant colony optimization
- computational cost
- search algorithm
- lower bound
- graph cuts
- particle filter
- evolutionary algorithm
- hidden markov models