Bayes Linear Statistics

Bayesian methods combine information available from data with anyprior information available from expert knowledge. The Bayes linearapproach follows this path, offering a quantitative structure forexpressing beliefs, and systematic methods for adjusting thesebeliefs, given observational data. The methodology differs from thefull Bayesian methodology in that it establishes simpler approachesto belief specification and analysis based around expectationjudgements. Bayes Linear Statistics presents anauthoritative account of this approach, explaining the foundations,theory, methodology, and practicalities of this important field. The text provides a thorough coverage of Bayes linear analysis,from the development of the basic language to the collection ofalgebraic results needed for efficient implementation, withdetailed practical examples. The book covers: * The importance of partial prior specifications for complexproblems where it is difficult to supply a meaningful full priorprobability specification. * Simple ways to use partial prior specifications to adjustbeliefs, given observations. * Interpretative and diagnostic tools to display the implicationsof collections of belief statements, and to make stringentcomparisons between expected and actual observations. * General approaches to statistical modelling based upon partialexchangeability judgements. * Bayes linear graphical models to represent and display partialbelief specifications, organize computations, and display theresults of analyses. Bayes Linear Statistics is essential reading for allstatisticians concerned with the theory and practice of Bayesianmethods. There is an accompanying website hosting free software andguides to the calculations within the book.