These are the things that I or other people have spotted. No doubt this list will continue to grow over time. Let me know if you find anything.
Bastian Blankenburg spotted a hideous inversion in the formula for Dempster's rule of combination. Rather than (in LaTeX notation):
1-\sum_{A\cap B = \emptyset}I have:
1-\sum_{A\cap B \neq \emptyset}Page 61.
Towards the bottom of the page I discuss some work by Paul Snow which sugegsts that it is possible to build a Dutch Book against the pignistic transformation from the Transferable Belief Model interpretation of Dempster Shafer theory. It turns out that I missed something here. The architect of the TBM, Philippe Smets, wrote to tell me:
One point about Snow's paper on Dutch Book. The problem had already been described and solved in a paper I published in 1993 : see my web site, look at:Smets Ph.(1993) No Dutch Book can be built against the TBM even though update is not obtained by Bayes rule of conditioning SIS, Workshop on Probabilistic Expert Systems. R. Scozzafava (ed.), Roma, Italy (1993) 181-204.
Page 259.
There is a typo in the set of equations following "and the second would mean that" towards the bottom of the page. The final symbol in the middle equation should be a [+]. Agonisingly I spotted this just after the manuscript had gone to press, so the version of these equations that appears on the front cover is correct...
S.D.Parsons@csc.liv.ac.uk |