|
| << vorige | volgende >> |
Tijdschrift beschrijving
Alle jaargangen van het bijbehorende tijdschrift
Alle afleveringen van het bijbehorende jaargang
Alle artikelen van de bijbehorende aflevering
Details van artikel 2 van 25 gevonden artikelen
| Titel: |
A Comparison of plotting rules under L1 and L2 estimation of the Weibull scale and shape parameters in situations of small samples with possible censoring and outliers |
| Auteur: |
Homan, Sharon M. |
| Verschenen in: |
Communications in statistics |
| Paginering: | Jaargang 18 (1989) nr. 1 pagina's 121-143 |
| Jaar: |
1989 |
| Inhoud: |
Certain biomedical applications of the Weibull distribution (e.g. time-to-relapse data of alcoholics) involve graphical estimation of the Weibull scale and shape parameters in small samples with possible outliers and potential censoring. This paper describes and compares the performance of six plotting conventions using least absolute deviation (L1) and least squares (L2) regression methods under conditions of small sample size, censoring, and outliers. A series of simulation experiments indicate that in small samples (N = 10), L2 regression methods are superior in all ten situations of censoring and contamination. When sample the size is small, the Bain and Antle plotting method is consistently the best choice in the absence of censoring, whereas the Kaplan-Meier plotting rule is consistently good when data are censored. For moderate size samples (N = 20), there is no clear relationship between L1 superiority with level and type of contamination. Further, the choice of plotting convention varies when data are uncensored. Again, the Kaplan-Meier plotting rule is clearly the better choice when data are censored. |
| Uitgever: |
Taylor & Francis |
| Bronbestand: |
Elektronische Wetenschappelijke Tijdschriften |
Details van artikel 2 van 25 gevonden artikelen
| << vorige | volgende >> |