I'm trying to calculate the shape and scale based on mean and standard deviation of a weibull distribution. If mean = 0 and sd = 1, the shape and scale both return NA. But for other values(such as mean = 1 and sd = 2), the result is perfect. Any requirement for the input parameters? I' m a beginner of statistic and R. Can someone help me out or give me any hint or reference? Thanks you so much.
mu = 0
sd = 1
sp = mixdist::weibullpar(mu, sd, loc = 0)$shape
sc = mixdist::weibullpar(mu, sd, loc = 0)$scale
asked Apr 10, 2023 at 19:57
Timon
3272 gold badges5 silver badges10 bronze badges
-
1What do you know about the Weibull distribution? In particular does it make any sense for a distribution that can only take values greater than 0 to have a mean of 0?Dason– Dason2023年04月10日 20:06:09 +00:00Commented Apr 10, 2023 at 20:06
-
2The support for Weibull distribution is 0 to positive infinite. It is impossible to have mean 0 and sd 1one– one2023年04月10日 20:08:06 +00:00Commented Apr 10, 2023 at 20:08
-
Hello Dason, thank you for the comment. I'm doing a simulation by creating a set of weibull distribution data based on mean and sd. So do you mean there is no practical significance passing mean = 0?Timon– Timon2023年04月10日 20:12:24 +00:00Commented Apr 10, 2023 at 20:12
-
Thank you peter861222, can you provide me some reference please?Timon– Timon2023年04月10日 20:13:23 +00:00Commented Apr 10, 2023 at 20:13
-
1@Timon Not only is there no practical significance - it's just impossible.Dason– Dason2023年04月10日 20:15:29 +00:00Commented Apr 10, 2023 at 20:15
1 Answer 1
For closure, the reason why a mean of 0 is inappropriate is that the Weibull distribution is defined on [0, ∞). Unless the distribution always returns 0—and is thus no longer a Weibull—the mean must be greater than 0.
Sign up to request clarification or add additional context in comments.
Comments
Explore related questions
See similar questions with these tags.
lang-r