Bayesian Forecasting and Dynamic Models (2nd edn)
- PDF / 205,865 Bytes
- 6 Pages / 595 x 842 pts (A4) Page_size
- 9 Downloads / 213 Views
#1998 Operational Research Society Ltd. All rights reserved. 0160-5682/98 $12.00
Book Selection Edited by JM Wilson M West and J Harrison: Bayesian Forecasting and Dynamic Models (2nd edn) L Oakshott: Business Modelling and Simulation EM Goldratt: Critical Chain AL Golub: Decision Analysis: An Integrated Approach S Cropper and P Forte (eds): Enhancing Health Services Management RS Barr, RV Helgason and JL Kennington (eds): Interfaces in Computer Science and Operations Research: Advances in Metaheuristics, Optimization, and Stochastic Modeling Technologies
Bayesian Forecasting and Dynamic Models (2nd edn) M West and J Harrison Springer, London, 1997. xiv 680 pp. DM 88.00. ISBN 0 387 94725 6 It puzzles me how slow the business community is to implement certain signi®cant ideas and yet how quickly management fads are readily adopted. Anyone who has been involved with sales forecasting will be aware of the insistence by managers of the need to incorporate business judgement into forecasts. Indeed, managers in some very well known manufacturing businesses that do have sophisticated computer systems for inventory control have been known to largely ignore or override their automated forecasts because they are unable to combine their judgement in a systematic way with the evidence of historical sales records. The key contribution of Bayesian forecasting is, in my view, the ability to combine judgement and historical evidence in a rigorous, meaningful manner. This major advance in forecasting practice has still to be realised in most businesses. For these reasons I am delighted to see that Bayesian Forecasting and Dynamic Models is now into its second edition and to have the opportunity of reviewing it. The book begins with essential notions of scienti®c method, modelling and Bayesian thinking. The authors ®rmly establish their `street credibility' as practising statisticians with several examples from real life that will ®nd echoes in the experience of all readers who have met practical forecasting problems. Though occupying only a small part of the book, these anecdotes contain much wisdom for people who have only been exposed to timeseries theory. They also show why the Bayesian dynamic
179 180 181 181 182 183
modelling approach is relevant. In the next two chapters the simplest models are introduced. Key mathematical results are highlighted, so that it is easy for the reader to see how to apply the methods. In fact, it is not dif®cult to transfer the relevant formulas to a spreadsheet and replicate the examples in the text at this stage. In the fourth chapter, the core ideas about dynamic linear models (DLMs) are elaborated. At this point, after the simplest models have been explained, the reader needs to be prepared for the mathematics that follows. The authors have provided a ®nal chapter, which is essentially an appendix, that brings together the mathematical prerequisites. These include elements of matrix algebra, the normal/ gamma distribution theory for Bayesian analyses and the normal theory related to linea
Data Loading...
