Methods for the Determination of the Probability Distribution and/or the Parameters Which Best Describe a Set of Test Data.

摘要

The purpose of this research is to analyze the methods for fitting standard probability density functions to test data. A logical method of combining the available techniques is then developed to offer a systematic data analysis procedure. The basic analysis techniques analyzed are the ones relating to the continuous density functions most often used in the fields of reliability and maintainability. Methods for fitting data investigated include the classical graphical techniques and the analytical methods of parameter estimation. Also, a wide range of useful computer applications is investigated. Two very useful tables,one a table of applications of statistical distributions and the other a table of maximum likelihood estimates,are included and serve as a quick reference. (Author)