## Applied Statistical Decision TheoryDivision of Research, Graduate School of Business Adminitration, Harvard University, 1961 - Business & Economics - 356 pages "In the field of statistical decision theory, Raiffa and Schlaifer have sought to develop new analytic techniques by which the modern theory of utility and subjective probability can actually be applied to the economic analysis of typical sampling problems." --From the foreword to their classic work "Applied Statistical Decision Theory," First published in the 1960s through Harvard University and MIT Press, the book is now offered in a new paperback edition from Wiley |

### From inside the book

Results 1-3 of 25

Page 125

...

...

**EVSI**as a Function of n In order to understand the behavior of the**EVSI**as a function of the sample size n , we first observe that as n increases from 0 through integral values the func- tion p defined by ( 5-51b ) at first retains some ...Page 129

...

...

**EVSI**as we have defined the**EVSI**in Section 4.5 . If on the contrary c is not optimal for the given n , i.e. , if cpm , then the value of the decision rule as given by ( 5-54 ) is less than the**EVSI**. If we were to plot the value of ( 5 ...Page 168

...

...

**EVSI**, we shall work out the arbitrarily chosen case where 10 observations are to be taken on each process and therefore ( since h ; / h = 1 for all i ) The first step is to compute n = 10 I. n " -1 = .05 I I , n " = n ' + n = 20 I , n ...### Contents

The Problem and the Two Basic Modes of Analysis | 3 |

Univariate Normalized Mass and Density Functions | 7 |

Combination of Formal and Informal Analysis | 17 |

Copyright | |

15 other sections not shown

### Other editions - View all

### Common terms and phrases

a₁ a₂ approximation assign assumption Bernoulli process beta function binomial choose compute conditional measure conjugate Conjugate prior cost cumulative function data-generating process decision maker decision problem decision tree defined definition denote estimate evaluated EVPI EVSI example expected terminal opportunity expected utility expected value experiment experimental outcome extensive form Ezle Figure follows gamma gamma-2 given h is known h is unknown k₁ k₂ kernel li(a li(e li(eo likelihood linear marginal measure mass function n₁ normalized density function observations obtain optimal act optimal sample parameter perfect information Poisson possible posterior density posterior distribution preposterior analysis prior density prior distribution prior expected probability quantity random variable sample information Section stopping process Substituting sufficient statistic Table terminal act terminal analysis terminal opportunity loss terminal utility theorem tion u₁ utility characteristic value of perfect vector vi(e