Dear all,
Sorry for cross mail.
I am using SAS 8.2 to do the goodness of fit test. The test is for 2
distributions: Lognormal and Weibull. It gives the KS test result for
lognormal. But it does not give the result for Weibull. Could anyone please
help me to point out the reason?
Thanks again in advance.
Regards
Wenqun Wang
ITS
University of Leeds
The result is as following:
1 21:00
Thursday, July 3, 2003 2
The UNIVARIATE Procedure
Fitted Distributions for V16
Parameters for Lognormal Distribution
Parameter Symbol Estimate
Threshold Theta -14.9467
Scale Zeta 4.152231
Shape Sigma 0.541327
Mean 58.66087
Std Dev 42.9511
Goodness-of-Fit Tests for Lognormal Distribution
Test ---Statistic---- -----p Value---
--
Kolmogorov-Smirnov D 0.07941122 Pr > D
>0.250
Cramer-von Mises W-Sq 0.03761332 Pr > W-Sq
>0.500
Anderson-Darling A-Sq 0.23392857 Pr > A-Sq
>0.500
Quantiles for Lognormal Distribution
--------Quantile-------
Percent Observed Estimated
1.0 3.00000 3.09913
5.0 8.00000 11.15046
10.0 18.00000 16.82233
25.0 31.00000 29.18203
50.0 50.00000 48.62895
75.0 75.00000 76.64586
90.0 120.00000 112.27979
95.0 133.00000 139.93073
99.0 238.00000 209.03069
Parameters for Weibull Distribution
Parameter Symbol Estimate
Threshold Theta 1.962913
Scale Sigma 61.82028
Shape C 1.407599
Mean 58.25994
Std Dev 40.54051
Quantiles for Weibull Distribution
--------Quantile-------
Percent Observed Estimated
1.0 3.00000 4.31699
5.0 8.00000 9.45702
10.0 18.00000 14.46010
25.0 31.00000 27.47386
50.0 50.00000 49.61135
75.0 75.00000 79.92963
90.0 120.00000 113.76751
95.0 133.00000 136.75139
99.0 238.00000 184.90780
|