Optimal Monte Carlo Estimation of Belief Network Inference
نویسندگان
چکیده
We present two Monte Carlo sampling algo rithms for probabilistic inference that guarantee polynomial-time convergence for a larger class of network than current sampling algorithms pro vide. These new methods are variants of the known likelihood weighting algorithm. We use of recent advances in the theory of optimal stopping rules for Monte Carlo simulation to obtain an inference approximation with relative error e and a small failure probability 5. We present an empirical evaluation of the algorithms which demonstrates their improved performance.
منابع مشابه
Applying Point Estimation and Monte Carlo Simulation Methods in Solving Probabilistic Optimal Power Flow Considering Renewable Energy Uncertainties
The increasing penetration of renewable energy results in changing the traditional power system planning and operation tools. As the generated power by the renewable energy resources are probabilistically changed, the certain power system analysis tolls cannot be applied in this case. Probabilistic optimal power flow is one of the most useful tools regarding the power system analysis in presen...
متن کاملOptimal Scheduling of Battery Energy Storage System in Distribution Network Considering Uncertainties using hybrid Monte Carlo- Genetic Approach
This paper proposes a novel hybrid Monte Carlo simulation-genetic approach (MCS-GA) for optimal operation of a distribution network considering renewable energy generation systems (REGSs) and battery energy storage systems (BESSs). The aim of this paper is to design an optimal charging /discharging scheduling of BESSs so that the total daily profit of distribution company (Disco) can be maximiz...
متن کاملBayesian Inference and Optimal Design in the Sparse Linear Model
The sparse linear model has seen many successful applications in Statistics, Machine Learning, and Computational Biology, such as identification of gene regulatory networks from micro-array expression data. Prior work has either approximated Bayesian inference by expensive Markov chain Monte Carlo, or replaced it by point estimation. We show how to obtain a good approximation to Bayesian analys...
متن کاملInference on Pr(X > Y ) Based on Record Values From the Power Hazard Rate Distribution
In this article, we consider the problem of estimating the stress-strength reliability $Pr (X > Y)$ based on upper record values when $X$ and $Y$ are two independent but not identically distributed random variables from the power hazard rate distribution with common scale parameter $k$. When the parameter $k$ is known, the maximum likelihood estimator (MLE), the approximate Bayes estimator and ...
متن کاملData-driven Sequential Monte Carlo in Probabilistic Programming
Most of Markov Chain Monte Carlo (MCMC) and sequential Monte Carlo (SMC) algorithms in existing probabilistic programming systems suboptimally use only model priors as proposal distributions. In this work, we describe an approach for training a discriminative model, namely a neural network, in order to approximate the optimal proposal by using posterior estimates from previous runs of inference...
متن کامل