Off-campus UNL users: To download campus access dissertations, please use the following link to log into our proxy server with your NU ID and password. When you are done browsing please remember to return to this page and log out.
Non-UNL users: Please talk to your librarian about requesting this dissertation through interlibrary loan.
BOOTSTRAP STATISTICS FOR EXPERT DATA INTERPRETATION (JACKKNIFE)
Abstract
A computer-intensive method for estimating small sample statistics and obtaining confidence intervals for them has been synthesized. This method combines the Jackknife and the Bootstrap methods. The combined method, called JBC (Jackknife and Bootstrap Combined), yields more reliable or smaller confidence intervals for a variety of statistical estimates. Computer experiments were designed to compare the results of the three methods: Jackknife, Bootstrap, and the JBC method. Programs implementing algorithms for each method were applied to both synthetically generated and real observed data. Comparison of results establishes that the JBC method is superior to the Jackknife and Bootstrap methods for obtaining smaller confidence intervals. This has been demonstrated with a variety of real and artificial data for the six statistics: mean, median, variance, skewness, kurtosis, and correlation coefficient. The improved results obtained by the JBC method are of value for analysis of small data samples and development of expert systems designed for writing statistical reports. A prototype of a text generation system for writing statistical reports has been developed. This system combines the results of Bootstrap method of statistical estimation with a small knowledge base consisting of rules. A "template-based" approach was used by a LISP pattern matching inference engine to produce appropriate English assertions.
Subject Area
Computer science
Recommended Citation
PIN-NGERN, OUEN, "BOOTSTRAP STATISTICS FOR EXPERT DATA INTERPRETATION (JACKKNIFE)" (1986). ETD collection for University of Nebraska-Lincoln. AAI8614471.
https://digitalcommons.unl.edu/dissertations/AAI8614471