IJPAM: Volume 47, No. 1 (2008)

Invited Lecture Delivered at
Forth International Conference of Applied Mathematics
and Computing (Plovdiv, Bulgaria, August 12-18, 2007)


Sorana D. Bolboaca$^1$, Lorentz Jäntschi$^2$
$^1$Medical Informatics and Biostatistics
``Iuliu Hatieganu" University of Medicine and Pharmacy
Cluj-Napoca, 400349, ROMANIA
e-mail: sorana@j.academicdirect.ro
$^2$Technical University of Cluj-Napoca
Cluj-Napoca, 400641, ROMANIA
e-mail: lori@academicdirect.org

Abstract.The aim of the research was to develop an optimization procedure of computing confidence intervals for binomial distributed samples based. An inductive algorithm stands as method used to solve the problem of confidence intervals estimation for binomial proportions. The implemented optimization procedure uses two triangulations (varying simultaneously two pairs of three variables). The optimization method was assessed in a simulation study for a significance level of 5%, and sample sizes that vary from six to one thousand and associated possible proportions. The obtained results are available online [#!l!#]. Overall, the optimization method performed better, the values of cumulative error function decreasing in average with 10%, depending on the sample sizes and the confidence intervals method with which it is compared. The performances of the optimization method increase toghether with sample size, surprisingly because it is well known that the confidence interval methods that use the normal approximation hypothesis for a binomial distribution obtain good results with increasing of sample sizes.

Received: August 17, 2007

AMS Subject Classification: 49M25, 60A05, 94B70, 62P10, 62HXX

Key Words and Phrases: optimization, confidence interval, binomial distribution, contingency table

Source: International Journal of Pure and Applied Mathematics
ISSN: 1311-8080
Year: 2008
Volume: 47
Issue: 1