Reassessing Accuracy Rates of Median Decisions

Capotorti, Andrea; Lad, Frank; Sanfilippo, Giuseppe

doi:10.1198/000313007X190943

by Andrea Capotorti, Frank Lad, Giuseppe Sanfilippo

Abstract:

We show how Bruno de Finetti's fundamental theorem of prevision has computable applications in statistical problems that involve only partial information. Specifically, we assess accuracy rates for median decision procedures used in the radiological diagnosis of asbestosis. Conditional exchangeability of individual radiologists' diagnoses is recognized as more appropriate than independence which is commonly presumed. The FTP yields coherent bounds on probabilities of interest when available information is insufficient to determine a complete distribution. Further assertions that are natural to the problem motivate a partial ordering of conditional probabilities, extending the computation from a linear to a quadratic programming problem.

View PDF

Reference:

Andrea Capotorti, Frank Lad, Giuseppe Sanfilippo, "Reassessing Accuracy Rates of Median Decisions", In The American Statistician, American Statistical Association, vol. 61, no. 2, pp. 132-138, 2007. ([A version similar to the published paper])

Bibtex Entry:

@ARTICLE{2007CLS-TAS,
  author = {Andrea Capotorti and Frank Lad and Giuseppe Sanfilippo},
  title = {Reassessing Accuracy Rates of Median Decisions},
  journal = {The American Statistician},
  year = {2007},
  volume = {61},
  pages = {132--138},
  number = {2},
  note = {ISSN: 0003-1305},
  issn={0003-1305},
  abstract = {We show how Bruno de Finetti's fundamental theorem of prevision has
	computable applications in statistical problems that involve only
	partial information. Specifically, we assess accuracy rates for median
	decision procedures used in the radiological diagnosis of asbestosis.
	Conditional exchangeability of individual radiologists' diagnoses
	is recognized as more appropriate than independence which is commonly
	presumed. The FTP yields coherent bounds on probabilities of interest
	when available information is insufficient to determine a complete
	distribution. Further assertions that are natural to the problem
	motivate a partial ordering of conditional probabilities, extending
	the computation from a linear to a quadratic programming problem.},
  comment = {<a href="http://sites.unipa.it/sanfilippo/pdf/2007/tas/2904_2_supp_1_j9k58u_cnvpdf.pdf">[A
	version similar to the published paper]</a>},
  doi = {10.1198/000313007X190943},
  mrnumber = {2368102},
  publisher = {American Statistical Association},
  scopus = {{2-s2.0-34248362401}},
  url = {http://dx.doi.org/10.1198/000313007X190943},
  wos = {{WOS:000245762200005}}
}