## Proceedings of the International School of Physics "Enrico Fermi", Volume 94 |

### From inside the book

Results 1-3 of 40

Page 150

A statistic s is sufficient for 0 if and only if , for every prior n ( 0 ) , the posterior a ( 0

| D ) = P ( DO ) # ( 0 ) / / p ( D | O ) = ( 0 ) do

the statistic s ; i . e . , for every prior a , the posterior can be written as a ( 0 s ) .

A statistic s is sufficient for 0 if and only if , for every prior n ( 0 ) , the posterior a ( 0

| D ) = P ( DO ) # ( 0 ) / / p ( D | O ) = ( 0 ) do

**depends**on the data D only throughthe statistic s ; i . e . , for every prior a , the posterior can be written as a ( 0 s ) .

Page 325

The size of an error

selection strategies , like boundary - value testing , path testing and range resting

, magnify the size of an error since they exercise error - prone constructs .

The size of an error

**depends**on the way the inputs are selected . Good test caseselection strategies , like boundary - value testing , path testing and range resting

, magnify the size of an error since they exercise error - prone constructs .

Page 384

The constant of proportionality , which

stopping rule , is not relevant . This matter regarding a specification of the

stopping rule , so fundamental to the use of sample theory statistics when applied

to inference ...

The constant of proportionality , which

**depends**on the ( noninformative )stopping rule , is not relevant . This matter regarding a specification of the

stopping rule , so fundamental to the use of sample theory statistics when applied

to inference ...

### What people are saying - Write a review

We haven't found any reviews in the usual places.

### Contents

STATISTICAL THEORY OF RELIABLITY | 8 |

Definitions and characterizations | 12 |

J KEILSON Stochastic models in reliability theory | 23 |

Copyright | |

37 other sections not shown

### Common terms and phrases

analysis application approach associated assume assumption BARLOW Bayesian calculation called complex components consider constant continuous correctness Course defined density depends derived described detected determine discussed distribution edited epochs equations equivalence ergodic errors estimate example exists expected exponential fact fail failure rate fault function given Hence important increasing independent input integration interest interval known likelihood limit Markov matrix mean measure method modules normal Note observed obtain occur operational parameters performance phase positive possible posterior predictive prior probability problem procedure prove random variables renewal repair requirements rule sample selected sequence simple software reliability space specification statistical stochastic structure Suppose task theorem theory tion transition tree University values York