 | | |
| Additive Representations of Preferences: A New Foundation of Decision Analysis (Theory and Decision Library C) | 
 enlarge | Author: P.p. Wakker
Publisher: Springer
Category: Book
List Price: $168.00
Buy New: $166.32
You Save: $1.68 (1%)
Buy New/Used from $166.32
Avg. Customer Rating:  (1 reviews)
Sales Rank: 3171583
Languages: English (Original Language), English (Unknown), English (Published)
Media: Hardcover
Edition: 1
Number Of Items: 1
Pages: 212
Shipping Weight (lbs): 1.1
Dimensions (in): 9.2 x 6.1 x 0.6
ISBN: 0792300505
Dewey Decimal Number: 658.4033
EAN: 9780792300502
ASIN: 0792300505
Publication Date: December 31, 1988
Availability: Usually ships in 1-2 business days
|
| Customer Reviews:
Keynes's foundation for decision analysis is superior December 15, 2004
0 out of 1 found this review helpful
Wakker(W)presents an interesting formal,mathematical,axiomatic treatment of a non- Subjective Expected Utility approach to decision making .W's intent is to provide a generalization of the SEU approach that can deal with Ellsberg-Popper type paradoxes,Allais problems and Tversky-Kahneman Prospect problems,as well as the standard type of risk problem where the probabilities are additive,linear,unique,definite single real numbers.The major criticism of the book is that it overlooks the contributions to the foundations of decision analysis made by John Maynard Keynes in 1921 in his A Treatise on Probability in chapters 6, 15,17,20,22,26,29,and 30.Keynes's system is much easier to understand,apply and use in the real world of actual everyday decision making where the probabilities are non-additive,ambiguous and unclear than the extremely cumbersome ,axiomatic approach of W.Keynes was the first writer on decision theory to specify a clearly defined mathematical approach to the determination of probabilistic interval(set) estimates.Keynes was the first to provide an axiomatic foundation that would apply to both numerical probabilities as well as "nonnumerical"probabilities(Keynes called his interval estimates nonnumerical probabilities .F P Ramsey completely misinterpreted Keynes's meaning and assumed that Keynes meant that numbers could not in general be used in the estimation of probabilities ,except in the special case of symmetrical evidence.This conclusion has misled economists,philosophers and decision theorists for over 80 years).Keynes was the first decision theorist to specify an index to measure the credibility of the evidential base upon which the probabilities would be calculated.Keynes defined the variable w to measure the weight of the evidence.w measured the completeness of the relevant actual and potential evidence available to a decision maker in order to estimate probabilities.w was defined by Keynes on the unit interval [0,1].Thus,0<=w<=1.Keynes was the first to provide a generalization of either the expected(monetary) value rule,maximize pA or the expected utility rule,maximize pU(A),by creating his conventional coefficient of risk and weight,c.The goal of the decision maker is to maximize cA,where c=p/(1+q)[2w/(1+w)].A modern name for Keynes's conventional coefficient would be to call it a decision weight that deals with ambiguity aversion.One need only invert the term [2w/(1+w)]to deal with ambiguity preference.Keynes's decision theory appeared before Wakker or D.Ellsberg was born.Keynes's w is a correspondence that is a one-to- one onto mapping with Ellsberg's rho.This mapping is isomorphic to Ellsberg's rho variable.Keynes's decision rule,maximize cA, is more general in some respects,overall,than Ellsberg's 1961 model in the range of its application,as the c coefficient also solves problems like the certainty effect,the reflection effect,the translation effect,the preference reversal effect,etc.Based on an application of Ockham's Razor,Keynes's approach is preferred to W's approach or Ellsberg's 1961 approach(Ellsberg's 1962 dissertation,published in 2001,is more general than Keynes's model in the relevant areas of application)in the opinion of this reviewer.
|
|
|
Powered by Associate-O-Matic
| | |