A Correlation Type Goodness of Fit Test for the Burr Distribution | ||||
| The Egyptian Statistical Journal | ||||
| Article 5, Volume 40, Issue 2, December 1996, Page 182-190 | ||||
| Document Type: Original Article | ||||
| DOI: 10.21608/esju.1996.314789 | ||||
 View on SCiNiTO | ||||
| Author | ||||
| Abd-Allah M. Abd-Elfattah | ||||
| Abstract | ||||
| The Burr family of distributions is widely used in life testing studies. The two common survival or failure-time distributions, the Weibull and the Exponential, are both special cases or limiting cases of the Burr type XII (see Lewis (1981), denoted by Burr (c,k), where c and k are the shape parameters. Goodness of fit tests are used customarily when the form of the population is in question, with the hope that the null hypothesis will be accepted. The compatibility of a set of observed sample values with any distribution can be checked by a goodness of fit test. These tests are designed for checking the validity of a null hypothesis, which is a statement about the form of the cumulative distribution function or probability function of the parent population from which the sample is drawn. One of the easiest goodness of fit tests to use is the "correlation coefficient" test. This simple test is easy to use, it requires a special table derived from Monte Carlo simulations. This test is Performed by ranking the data, associating with each datum the expected value of the order statistic with the same rank, computing the product-moment correlation coefficient between the data and the Burr deviates, and finally, using a table to find the probability of a good fit associated with the observed correlation. Calculated correlations that are smaller than the critical value indicate lack of fit. | ||||
| Keywords | ||||
| Burr Distribution; Correlation Coefficient; Goodness of Fit Tests; Monte Carlo Simulation; Critical Values | ||||
|
Statistics Article View: 21 |
||||
