A fast-weighted Bayesian bootstrap filter for nonlinear model state estimation

Beadle, E.R.; Djurić, Petar M.

doi:10.1109/7.570818

Cited by 73 publications

(38 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This approach is the core of many successful sequential Bayesian tools, such as the well-known bootstrap or particle filters (Avitzour, 1995;Beadle & Djuric, 1997;Berzuini et al, 1997;Doucet, 1998;Gordon et al, 1993;Kitagawa, 1996;Liu & Chen, 1998;Pitt & Shephard, 1997). In the following sections, we derive this algorithm from an importance sampling perspective.…”

Section: (I)mentioning

confidence: 99%

Sequential Monte Carlo Methods to Train Neural Network Models

et al. 2000

View full text Add to dashboard Cite

We discuss a novel strategy for training neural networks using sequential Monte Carlo algorithms and propose a new hybrid gradient descent/ sampling importance resampling algorithm (HySIR). In terms of computational time and accuracy, the hybrid SIR is a clear improvement over conventional sequential Monte Carlo techniques. The new algorithm may be viewed as a global optimization strategy that allows us to learn the probability distributions of the network weights and outputs in a sequential framework. It is well suited to applications involving on-line, nonlinear, and nongaussian signal processing. We show how the new algorithm outperforms extended Kalman filter training on several problems. In particular, we address the problem of pricing option contracts, traded in financial markets. In this context, we are able to estimate the one-step-ahead probability density functions of the options prices.

show abstract

Section: (I)mentioning

confidence: 99%

Sequential Monte Carlo Methods to Train Neural Network Models

et al. 2000

View full text Add to dashboard Cite

show abstract

“…The bootstrap, on the other hand, guarantees each prior sample a chance, no matter how small, of being selected in the posterior. To overcome this problem, two ways were suggested in [17]. One is to choose an m large enough so that [mq i ] ¿ 1 for almost every x * k (i) in the prior.…”

Section: Discussionmentioning

confidence: 99%

“…A fast bootstrap algorithm was proposed by Beadle and Djuric [17] to overcome the problem. It is based on the expected number of times at which a prior sample appears in the posterior at each time instant, as described below.…”

Section: Fast Bootstrap Techniquementioning

confidence: 99%

“…In this section, both the original bootstrap ÿlter [10] and the fast Bayesian bootstrap ÿlter [17] will be reviewed brie y.…”

Section: Fast Bayesian Bootstrap ÿLtermentioning

confidence: 99%

“…However, a drawback of the method is the heavy computational load, which precludes its utilization for real time estimations. A fast Bayesian bootstrap ÿl-ter, which can dramatically decreases the computation time of the standard bootstrap ÿlter, was proposed by Beadle and Djuric [17]. The time decrease is realized by resampling the prior into the posterior distribution at each time step by using sampling groups, rather than a sample as in the standard approach.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Improvement of parameter estimation for non-linear hysteretic systems with slip by a fast Bayesian bootstrap filter

Suzuki

Noori

2004

International Journal of Non-Linear Mechanics

View full text Add to dashboard Cite

Modeling and identiÿcation of non-linear hysteretic systems are widely encountered in the structural dynamics ÿeld, especially for the hysteresis with slip. A model, called SL model, which can describe the pinching of most practical hysteresis loops perfectly was proposed by Baber and Noori (J. Eng. Mech. 111 (1985) 1010). A method of estimating the parameters of SL model on the basis of input-output data based on bootstrap ÿlter was proposed by the writers. Bootstrap ÿlter is a ÿltering method based on Bayesian state estimation and Monte Carlo method, which has the great advantage of being able to handle any functional non-linearity and system and/or measurement noise of any distribution. The standard bootstrap ÿlter, however, is not time e cient, i.e., it is very time consuming and is not suitable for real-time applications. In this paper, previous work by the writers is extended to do the parameter estimation of SL model by a fast Bayesian bootstrap ÿltering technique. Simulation results are presented to demonstrate the performance of the algorithm.

show abstract

Kalman Filtering and Neural Networks

Haykin¹

2001

1,226

View full text Add to dashboard Cite

A fast-weighted Bayesian bootstrap filter for nonlinear model state estimation

Cited by 73 publications

References 4 publications

Sequential Monte Carlo Methods to Train Neural Network Models

Sequential Monte Carlo Methods to Train Neural Network Models

Improvement of parameter estimation for non-linear hysteretic systems with slip by a fast Bayesian bootstrap filter

Kalman Filtering and Neural Networks

Contact Info

Product

Resources

About