Learning in graphical models：（学习在图形化的模型）.pdfVIP

Statistical Science 2004, Vol. 19, No. 1, 140–155 DOI 10.1214/088342304000000026 © Institute of Mathematical Statistics, 2004 Graphical Models Michael I. Jordan Abstract. Statistical applications in fields such as bioinformatics, informa- tion retrieval, speech processing, image processing and communications of- ten involve large-scale models in which thousands or millions of random variables are linked in complex ways. Graphical models provide a gen- eral methodology for approaching these problems, and indeed many of the models developed by researchers in these applied fields are instances of the general graphical model formalism. We review some of the basic ideas under- lying graphical models, including the algorithmic ideas that allow graphical models to be deployed in large-scale data analysis problems. We also present examples of graphical models in bioinformatics, error-control coding and language processing. Key words and phrases: Probabilistic graphical models, junction tree algorithm, sum-product algorithm, Markov chain Monte Carlo, variational inference, bioinformatics, error-control coding. 1. INTRODUCTION representation, the formalism provides general algo- rithms for computing marginal and conditional prob- The fields of statistics and computer science have abilities of interest. Moreover, the formalism provides generally followed separate paths for the past several control