[Scilab-users] needed: SciLab routine for generating Weibull random variables with prior given autocorrelation
Heinz Nabielek
heinznabielek at me.com
Thu May 3 18:30:51 CEST 2018
thanks.
but still: how construct autocorrelated uniform randoms?
Sent from Heinz Nabielek
> On 03 May 2018, at 17:40, Stéphane Mottelet <stephane.mottelet at utc.fr> wrote:
>
>> Le 03/05/2018 à 15:49, Heinz Nabielek a écrit :
>> Many thanks. <https://antispam.utc.fr/proxy/2/c3RlcGhhbmUubW90dGVsZXRAdXRjLmZy/en.wikipedia.org/wiki/Autoregressive_model#Example:_An_AR(1)_process> works fine on an originating normal distribution [where i would not need it, because we have grand(n, "mn", Mean, Cov) in SciLab].
> grand(n, "mn", Mean, Cov) allows to generate realizations of a multi-normal variable, i.e. covariance matrix describes the covariance between each component of the (vector) variable.
>
> AR processes are different stuff (random process <> random variable). For an AR process the non-idependance is w.r.t. to time (discrete or continuous)
>
> Your wind data is a random process.
>> With an initiating U(-0.5, 0,5), however, I get a triangle function, but not a autocorrelated uniform random distribution.
> what do you mean by " autocorrelated uniform random distribution" ? Here your triangle is just the distribution of a sum of two uniform vraiables.
>
> S.
>> What to do?
>> Heinz
>>
>>
>>
>>
>>> On 03.05.2018, at 11:37, Stéphane Mottelet <stephane.mottelet at utc.fr> wrote:
>>>
>>> Why don't start with a simple autoregressive process ? See e.g.
>>>
>>> https://antispam.utc.fr/proxy/2/c3RlcGhhbmUubW90dGVsZXRAdXRjLmZy/en.wikipedia.org/wiki/Autoregressive_model
>>>
>>>
>>>> Le 03/05/2018 à 11:15, Heinz Nabielek a écrit :
>>>> Uniform U(0,1) random deviates with given autocorrelation would be good enough…...
>>>>
>>>>
>>>>> On 03.05.2018, at 10:57, Heinz Nabielek <heinznabielek at me.com> wrote:
>>>>>
>>>>> Friends:
>>>>>
>>>>> urgently needed: a SciLab routine for generating Weibull random variables with prior given autocorrelation.
>>>>>
>>>>> Uncorrelated is easy with
>>>>> vc*((-log(grand(1,5,"def"))).^(1/m)) producing
>>>>> 0.9666848 1.4646143 10.402793 5.9513675 3.3241823
>>>>> with characteristic velocity vc=7 and modulus m=2.
>>>>>
>>>>> But I need Weibull deviates with given autocorrelation, say rho=0.80
>>>>>
>>>>> Literature has dozen of references on the subject which either do not work or I do not know how to code.
>>>>>
>>>>> Any help here?
>>>>> Heinz
>> _______________________________________________
>> users mailing list
>> users at lists.scilab.org
>> https://antispam.utc.fr/proxy/1/c3RlcGhhbmUubW90dGVsZXRAdXRjLmZy/lists.scilab.org/mailman/listinfo/users
>
> _______________________________________________
> users mailing list
> users at lists.scilab.org
> http://lists.scilab.org/mailman/listinfo/users
More information about the users
mailing list