An introduction to Bayesian Networks and the Bayes Net Toolbox for Matlab课件.ppt

An introduction to Bayesian Networksand the Bayes Net Toolboxfor Matlab;Outline;Qualitative part: Directed acyclic graph (DAG) Nodes - random vars. Edges - direct influence;What is a Bayes net?;Why are Bayes nets useful?;What can Bayes nets be used for?;Domain: Monitoring Intensive-Care Patients 37 variables 509 parameters …instead of 237;More real-world BN applications;Dealing with time;State-space model (SSM)/Linear Dynamical System (LDS);Example: LDS for 2D tracking;Hidden Markov model (HMM);Probabilistic graphical models;Toy example of a Markov net;A real Markov net;Inference;Kalman filtering (recursive state estimation in an LDS) ;Forwards algorithm for HMMs;Message passing view of forwards algorithm;Forwards-backwards algorithm;Belief Propagation;BP: parallel, distributed version;Representing potentials;Manipulating discrete potentials;Manipulating Gaussian potentials;Semi-rings;Inference in general graphs;Approximate inference;Learning;Parameter learning;Structure learning (data mining);Structure learning;Summary:why are graphical models useful?;The Bayes Net Toolbox for Matlab;What is BNT?;Why yet another BN toolbox?;A comparison of GM software;Summary of existing GM software;Why Matlab?;BNT’s class structure;Example: mixture of experts;1. Making the graph;2. Making the model;3. Specifying the parameters;4. Training the model;Before training;After training;5. Inference/ prediction;Other kinds of models that BNT supports;Summary of BNT;What is wrong with BNT?;Some alternatives to BNT?;Why yet another GM toolbox?