Extreme value distribution

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
11 messages Options
Reply | Threaded
Open this post in threaded view
|

Extreme value distribution

David Anthoff

Hi,

 

is there a package dealing with extreme value distribution stuff? A quick search on pkg.julialang.org didn’t reveal one. I’m mainly looking for code that estimates the tail index of an empirical data set I have.

 

Thanks,

David

 

--

David Anthoff

University of California, Berkeley

 

http://www.david-anthoff.com

 

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.
Reply | Threaded
Open this post in threaded view
|

Re: Extreme value distribution

Andreas Noack
Hi David,

Distributions.jl has the Gumpel, Fréchet and Weibull distributions aka the type I, II and III extreme value distribution.

On Wed, Aug 12, 2015 at 5:25 PM, David Anthoff <[hidden email]> wrote:

Hi,

 

is there a package dealing with extreme value distribution stuff? A quick search on pkg.julialang.org didn’t reveal one. I’m mainly looking for code that estimates the tail index of an empirical data set I have.

 

Thanks,

David

 

--

David Anthoff

University of California, Berkeley

 

http://www.david-anthoff.com

 

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.
Reply | Threaded
Open this post in threaded view
|

RE: Extreme value distribution

David Anthoff

But no support for fitting them, right? Thanks, David

 

From: [hidden email] [mailto:[hidden email]] On Behalf Of Andreas Noack
Sent: Thursday, August 13, 2015 6:30 AM
To: [hidden email]
Subject: Re: [julia-stats] Extreme value distribution

 

Hi David,

 

Distributions.jl has the Gumpel, Fréchet and Weibull distributions aka the type I, II and III extreme value distribution.

 

On Wed, Aug 12, 2015 at 5:25 PM, David Anthoff <[hidden email]> wrote:

Hi,

 

is there a package dealing with extreme value distribution stuff? A quick search on pkg.julialang.org didn’t reveal one. I’m mainly looking for code that estimates the tail index of an empirical data set I have.

 

Thanks,

David

 

--

David Anthoff

University of California, Berkeley

 

http://www.david-anthoff.com

 

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to
[hidden email].
For more options, visit
https://groups.google.com/d/optout.

 

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.
Reply | Threaded
Open this post in threaded view
|

Re: Extreme value distribution

Simon Byrne
Not at the moment (as far as I know). If you don't want to implement it yourself, you could try RCall.jl.

On Thursday, 13 August 2015 18:09:22 UTC+1, David Anthoff wrote:

But no support for fitting them, right? Thanks, David

 

From: <a href="javascript:" target="_blank" gdf-obfuscated-mailto="CNIp73g3BAAJ" rel="nofollow" onmousedown="this.href=&#39;javascript:&#39;;return true;" onclick="this.href=&#39;javascript:&#39;;return true;">julia...@... [mailto:<a href="javascript:" target="_blank" gdf-obfuscated-mailto="CNIp73g3BAAJ" rel="nofollow" onmousedown="this.href=&#39;javascript:&#39;;return true;" onclick="this.href=&#39;javascript:&#39;;return true;">julia...@googlegroups.com] On Behalf Of Andreas Noack
Sent: Thursday, August 13, 2015 6:30 AM
To: <a href="javascript:" target="_blank" gdf-obfuscated-mailto="CNIp73g3BAAJ" rel="nofollow" onmousedown="this.href=&#39;javascript:&#39;;return true;" onclick="this.href=&#39;javascript:&#39;;return true;">julia...@...
Subject: Re: [julia-stats] Extreme value distribution

 

Hi David,

 

Distributions.jl has the Gumpel, Fréchet and Weibull distributions aka the type I, II and III extreme value distribution.

 

On Wed, Aug 12, 2015 at 5:25 PM, David Anthoff <<a href="javascript:" target="_blank" gdf-obfuscated-mailto="CNIp73g3BAAJ" rel="nofollow" onmousedown="this.href=&#39;javascript:&#39;;return true;" onclick="this.href=&#39;javascript:&#39;;return true;">ant...@...> wrote:

Hi,

 

is there a package dealing with extreme value distribution stuff? A quick search on <a href="http://pkg.julialang.org" target="_blank" rel="nofollow" onmousedown="this.href=&#39;http://www.google.com/url?q\75http%3A%2F%2Fpkg.julialang.org\46sa\75D\46sntz\0751\46usg\75AFQjCNH322kxbFE48sND8uB96ZIEXvdoCQ&#39;;return true;" onclick="this.href=&#39;http://www.google.com/url?q\75http%3A%2F%2Fpkg.julialang.org\46sa\75D\46sntz\0751\46usg\75AFQjCNH322kxbFE48sND8uB96ZIEXvdoCQ&#39;;return true;">pkg.julialang.org didn’t reveal one. I’m mainly looking for code that estimates the tail index of an empirical data set I have.

 

Thanks,

David

 

--

David Anthoff

University of California, Berkeley

 

<a href="http://www.david-anthoff.com" target="_blank" rel="nofollow" onmousedown="this.href=&#39;http://www.google.com/url?q\75http%3A%2F%2Fwww.david-anthoff.com\46sa\75D\46sntz\0751\46usg\75AFQjCNHPDijM7QgcCRbDWt_-VR5r_HnDGg&#39;;return true;" onclick="this.href=&#39;http://www.google.com/url?q\75http%3A%2F%2Fwww.david-anthoff.com\46sa\75D\46sntz\0751\46usg\75AFQjCNHPDijM7QgcCRbDWt_-VR5r_HnDGg&#39;;return true;">http://www.david-anthoff.com

 

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to
<a href="javascript:" target="_blank" gdf-obfuscated-mailto="CNIp73g3BAAJ" rel="nofollow" onmousedown="this.href=&#39;javascript:&#39;;return true;" onclick="this.href=&#39;javascript:&#39;;return true;">julia-stats...@googlegroups.com.
For more options, visit
<a href="https://groups.google.com/d/optout" target="_blank" rel="nofollow" onmousedown="this.href=&#39;https://groups.google.com/d/optout&#39;;return true;" onclick="this.href=&#39;https://groups.google.com/d/optout&#39;;return true;">https://groups.google.com/d/optout.

 

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to <a href="javascript:" target="_blank" gdf-obfuscated-mailto="CNIp73g3BAAJ" rel="nofollow" onmousedown="this.href=&#39;javascript:&#39;;return true;" onclick="this.href=&#39;javascript:&#39;;return true;">julia-stats...@googlegroups.com.
For more options, visit <a href="https://groups.google.com/d/optout" target="_blank" rel="nofollow" onmousedown="this.href=&#39;https://groups.google.com/d/optout&#39;;return true;" onclick="this.href=&#39;https://groups.google.com/d/optout&#39;;return true;">https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.
Reply | Threaded
Open this post in threaded view
|

RE: Extreme value distribution

David Anthoff

I coded up a very rough implementation of the Hill (1975) estimator, and then also another method by Huisman et al. (2001). Both methods are short (~10 lines), but they might be the start of a ExtremeValueTheory package.

 

The whole field is not central to what I do, so I would be the wrong person to maintain such a package, but I’d be happy to start it with the two estimators I coded up.

 

Should I just create an unregistered package with name ExtremeValueTheory that has the two algorithms? But people would probably never find it… Or would it make sense to have that package under JuliaStats from the beginning? Or is there some other way to share such one-off implementations with the community?

 

Cheers,

David

 

From: [hidden email] [mailto:[hidden email]] On Behalf Of Simon Byrne
Sent: Friday, August 14, 2015 1:24 AM
To: julia-stats <[hidden email]>
Subject: Re: [julia-stats] Extreme value distribution

 

Not at the moment (as far as I know). If you don't want to implement it yourself, you could try RCall.jl.

On Thursday, 13 August 2015 18:09:22 UTC+1, David Anthoff wrote:

But no support for fitting them, right? Thanks, David

 

From: <a href="javascript:" target="_blank">julia...@... [mailto:<a href="javascript:" target="_blank">julia...@...] On Behalf Of Andreas Noack
Sent: Thursday, August 13, 2015 6:30 AM
To:
<a href="javascript:" target="_blank">julia...@...
Subject: Re: [julia-stats] Extreme value distribution

 

Hi David,

 

Distributions.jl has the Gumpel, Fréchet and Weibull distributions aka the type I, II and III extreme value distribution.

 

On Wed, Aug 12, 2015 at 5:25 PM, David Anthoff <<a href="javascript:" target="_blank">ant...@...> wrote:

Hi,

 

is there a package dealing with extreme value distribution stuff? A quick search on pkg.julialang.org didn’t reveal one. I’m mainly looking for code that estimates the tail index of an empirical data set I have.

 

Thanks,

David

 

--

David Anthoff

University of California, Berkeley

 

http://www.david-anthoff.com

 

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to
<a href="javascript:" target="_blank">julia-stats...@....
For more options, visit
https://groups.google.com/d/optout.

 

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to <a href="javascript:" target="_blank">julia-stats...@....
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.
Reply | Threaded
Open this post in threaded view
|

Re: Extreme value distribution

John Myles White
If something is really one-off, I would just create a GitHub gist.

 -- John

On Aug 14, 2015, at 10:19 AM, David Anthoff <[hidden email]> wrote:

I coded up a very rough implementation of the Hill (1975) estimator, and then also another method by Huisman et al. (2001). Both methods are short (~10 lines), but they might be the start of a ExtremeValueTheory package.
 
The whole field is not central to what I do, so I would be the wrong person to maintain such a package, but I’d be happy to start it with the two estimators I coded up.
 
Should I just create an unregistered package with name ExtremeValueTheory that has the two algorithms? But people would probably never find it… Or would it make sense to have that package under JuliaStats from the beginning? Or is there some other way to share such one-off implementations with the community?
 
Cheers,
David
 
From: [hidden email] [[hidden email]] On Behalf Of Simon Byrne
Sent: Friday, August 14, 2015 1:24 AM
To: julia-stats <[hidden email]>
Subject: Re: [julia-stats] Extreme value distribution
 
Not at the moment (as far as I know). If you don't want to implement it yourself, you could try RCall.jl.

On Thursday, 13 August 2015 18:09:22 UTC+1, David Anthoff wrote:
But no support for fitting them, right? Thanks, David
 
From: [hidden email] [mailto:[hidden email]] On Behalf Of Andreas Noack
Sent: Thursday, August 13, 2015 6:30 AM
To: 
[hidden email]
Subject: Re: [julia-stats] Extreme value distribution
 
Hi David,
 
Distributions.jl has the Gumpel, Fréchet and Weibull distributions aka the type I, II and III extreme value distribution.
 
On Wed, Aug 12, 2015 at 5:25 PM, David Anthoff <[hidden email]> wrote:
Hi,
 
is there a package dealing with extreme value distribution stuff? A quick search on pkg.julialang.org didn’t reveal one. I’m mainly looking for code that estimates the tail index of an empirical data set I have.
 
Thanks,
David
 
--
David Anthoff
University of California, Berkeley
 
 
-- 
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to 
[hidden email].
For more options, visit 
https://groups.google.com/d/optout.
 
-- 
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.
-- 
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.

-- 
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.
Reply | Threaded
Open this post in threaded view
|

Re: Extreme value distribution

Patrick Kofod Mogensen
In reply to this post by David Anthoff
In what way is it one-off? If it works and is correct, why not PR it to Distributions.jl?

On Friday, August 14, 2015 at 7:19:44 PM UTC+2, David Anthoff wrote:

I coded up a very rough implementation of the Hill (1975) estimator, and then also another method by Huisman et al. (2001). Both methods are short (~10 lines), but they might be the start of a ExtremeValueTheory package.

 

The whole field is not central to what I do, so I would be the wrong person to maintain such a package, but I’d be happy to start it with the two estimators I coded up.

 

Should I just create an unregistered package with name ExtremeValueTheory that has the two algorithms? But people would probably never find it… Or would it make sense to have that package under JuliaStats from the beginning? Or is there some other way to share such one-off implementations with the community?

 

Cheers,

David

 

From: <a href="javascript:" target="_blank" gdf-obfuscated-mailto="BTkmOJ6GBAAJ" rel="nofollow" onmousedown="this.href=&#39;javascript:&#39;;return true;" onclick="this.href=&#39;javascript:&#39;;return true;">julia...@... [mailto:<a href="javascript:" target="_blank" gdf-obfuscated-mailto="BTkmOJ6GBAAJ" rel="nofollow" onmousedown="this.href=&#39;javascript:&#39;;return true;" onclick="this.href=&#39;javascript:&#39;;return true;">julia...@googlegroups.com] On Behalf Of Simon Byrne
Sent: Friday, August 14, 2015 1:24 AM
To: julia-stats <<a href="javascript:" target="_blank" gdf-obfuscated-mailto="BTkmOJ6GBAAJ" rel="nofollow" onmousedown="this.href=&#39;javascript:&#39;;return true;" onclick="this.href=&#39;javascript:&#39;;return true;">julia...@...>
Subject: Re: [julia-stats] Extreme value distribution

 

Not at the moment (as far as I know). If you don't want to implement it yourself, you could try RCall.jl.

On Thursday, 13 August 2015 18:09:22 UTC+1, David Anthoff wrote:

But no support for fitting them, right? Thanks, David

 

From: [hidden email] [mailto:julia...@googlegroups.com] On Behalf Of Andreas Noack
Sent: Thursday, August 13, 2015 6:30 AM
To:
[hidden email]
Subject: Re: [julia-stats] Extreme value distribution

 

Hi David,

 

Distributions.jl has the Gumpel, Fréchet and Weibull distributions aka the type I, II and III extreme value distribution.

 

On Wed, Aug 12, 2015 at 5:25 PM, David Anthoff <[hidden email]> wrote:

Hi,

 

is there a package dealing with extreme value distribution stuff? A quick search on <a href="http://pkg.julialang.org" target="_blank" rel="nofollow" onmousedown="this.href=&#39;http://www.google.com/url?q\75http%3A%2F%2Fpkg.julialang.org\46sa\75D\46sntz\0751\46usg\75AFQjCNH322kxbFE48sND8uB96ZIEXvdoCQ&#39;;return true;" onclick="this.href=&#39;http://www.google.com/url?q\75http%3A%2F%2Fpkg.julialang.org\46sa\75D\46sntz\0751\46usg\75AFQjCNH322kxbFE48sND8uB96ZIEXvdoCQ&#39;;return true;">pkg.julialang.org didn’t reveal one. I’m mainly looking for code that estimates the tail index of an empirical data set I have.

 

Thanks,

David

 

--

David Anthoff

University of California, Berkeley

 

<a href="http://www.david-anthoff.com" target="_blank" rel="nofollow" onmousedown="this.href=&#39;http://www.google.com/url?q\75http%3A%2F%2Fwww.david-anthoff.com\46sa\75D\46sntz\0751\46usg\75AFQjCNHPDijM7QgcCRbDWt_-VR5r_HnDGg&#39;;return true;" onclick="this.href=&#39;http://www.google.com/url?q\75http%3A%2F%2Fwww.david-anthoff.com\46sa\75D\46sntz\0751\46usg\75AFQjCNHPDijM7QgcCRbDWt_-VR5r_HnDGg&#39;;return true;">http://www.david-anthoff.com

 

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to
julia-stats...@googlegroups.com.
For more options, visit
<a href="https://groups.google.com/d/optout" target="_blank" rel="nofollow" onmousedown="this.href=&#39;https://groups.google.com/d/optout&#39;;return true;" onclick="this.href=&#39;https://groups.google.com/d/optout&#39;;return true;">https://groups.google.com/d/optout.

 

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to julia-stats...@googlegroups.com.
For more options, visit <a href="https://groups.google.com/d/optout" target="_blank" rel="nofollow" onmousedown="this.href=&#39;https://groups.google.com/d/optout&#39;;return true;" onclick="this.href=&#39;https://groups.google.com/d/optout&#39;;return true;">https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to <a href="javascript:" target="_blank" gdf-obfuscated-mailto="BTkmOJ6GBAAJ" rel="nofollow" onmousedown="this.href=&#39;javascript:&#39;;return true;" onclick="this.href=&#39;javascript:&#39;;return true;">julia-stats...@googlegroups.com.
For more options, visit <a href="https://groups.google.com/d/optout" target="_blank" rel="nofollow" onmousedown="this.href=&#39;https://groups.google.com/d/optout&#39;;return true;" onclick="this.href=&#39;https://groups.google.com/d/optout&#39;;return true;">https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.
Reply | Threaded
Open this post in threaded view
|

Re: Extreme value distribution

John Myles White
It's one-off because David has said he doesn't time to maintain it. One-off code can be great code, but even the best code eventually bit rots without active maintenace.

 -- John

On Aug 25, 2015, at 8:07 AM, Patrick Kofod Mogensen <[hidden email]> wrote:

In what way is it one-off? If it works and is correct, why not PR it to Distributions.jl?

On Friday, August 14, 2015 at 7:19:44 PM UTC+2, David Anthoff wrote:

I coded up a very rough implementation of the Hill (1975) estimator, and then also another method by Huisman et al. (2001). Both methods are short (~10 lines), but they might be the start of a ExtremeValueTheory package.

 

The whole field is not central to what I do, so I would be the wrong person to maintain such a package, but I’d be happy to start it with the two estimators I coded up.

 

Should I just create an unregistered package with name ExtremeValueTheory that has the two algorithms? But people would probably never find it… Or would it make sense to have that package under JuliaStats from the beginning? Or is there some other way to share such one-off implementations with the community?

 

Cheers,

David

 

From: <a href="javascript:" target="_blank" gdf-obfuscated-mailto="BTkmOJ6GBAAJ" rel="nofollow" onmousedown="this.href='javascript:';return true;" onclick="this.href='javascript:';return true;" class="">julia...@... [mailto:<a href="javascript:" target="_blank" gdf-obfuscated-mailto="BTkmOJ6GBAAJ" rel="nofollow" onmousedown="this.href='javascript:';return true;" onclick="this.href='javascript:';return true;" class="">julia...@googlegroups.com] On Behalf Of Simon Byrne
Sent: Friday, August 14, 2015 1:24 AM
To: julia-stats <<a href="javascript:" target="_blank" gdf-obfuscated-mailto="BTkmOJ6GBAAJ" rel="nofollow" onmousedown="this.href='javascript:';return true;" onclick="this.href='javascript:';return true;" class="">julia...@...>
Subject: Re: [julia-stats] Extreme value distribution

 

Not at the moment (as far as I know). If you don't want to implement it yourself, you could try RCall.jl.

On Thursday, 13 August 2015 18:09:22 UTC+1, David Anthoff wrote:

But no support for fitting them, right? Thanks, David

 

From: [hidden email] [mailto:julia...@googlegroups.com] On Behalf Of Andreas Noack
Sent: Thursday, August 13, 2015 6:30 AM
To:
[hidden email]
Subject: Re: [julia-stats] Extreme value distribution

 

Hi David,

 

Distributions.jl has the Gumpel, Fréchet and Weibull distributions aka the type I, II and III extreme value distribution.

 

On Wed, Aug 12, 2015 at 5:25 PM, David Anthoff <[hidden email]> wrote:

Hi,

 

is there a package dealing with extreme value distribution stuff? A quick search on <a href="http://pkg.julialang.org/" target="_blank" rel="nofollow" onmousedown="this.href='http://www.google.com/url?q\75http%3A%2F%2Fpkg.julialang.org\46sa\75D\46sntz\0751\46usg\75AFQjCNH322kxbFE48sND8uB96ZIEXvdoCQ';return true;" onclick="this.href='http://www.google.com/url?q\75http%3A%2F%2Fpkg.julialang.org\46sa\75D\46sntz\0751\46usg\75AFQjCNH322kxbFE48sND8uB96ZIEXvdoCQ';return true;" class="">pkg.julialang.org didn’t reveal one. I’m mainly looking for code that estimates the tail index of an empirical data set I have.

 

Thanks,

David

 

--

David Anthoff

University of California, Berkeley

 

<a href="http://www.david-anthoff.com/" target="_blank" rel="nofollow" onmousedown="this.href='http://www.google.com/url?q\75http%3A%2F%2Fwww.david-anthoff.com\46sa\75D\46sntz\0751\46usg\75AFQjCNHPDijM7QgcCRbDWt_-VR5r_HnDGg';return true;" onclick="this.href='http://www.google.com/url?q\75http%3A%2F%2Fwww.david-anthoff.com\46sa\75D\46sntz\0751\46usg\75AFQjCNHPDijM7QgcCRbDWt_-VR5r_HnDGg';return true;" class="">http://www.david-anthoff.com

 

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to
julia-stats...@googlegroups.com.
For more options, visit
<a href="https://groups.google.com/d/optout" target="_blank" rel="nofollow" onmousedown="this.href='https://groups.google.com/d/optout';return true;" onclick="this.href='https://groups.google.com/d/optout';return true;" class="">https://groups.google.com/d/optout.

 

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to julia-stats...@googlegroups.com.
For more options, visit <a href="https://groups.google.com/d/optout" target="_blank" rel="nofollow" onmousedown="this.href='https://groups.google.com/d/optout';return true;" onclick="this.href='https://groups.google.com/d/optout';return true;" class="">https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to <a href="javascript:" target="_blank" gdf-obfuscated-mailto="BTkmOJ6GBAAJ" rel="nofollow" onmousedown="this.href='javascript:';return true;" onclick="this.href='javascript:';return true;" class="">julia-stats...@googlegroups.com.
For more options, visit <a href="https://groups.google.com/d/optout" target="_blank" rel="nofollow" onmousedown="this.href='https://groups.google.com/d/optout';return true;" onclick="this.href='https://groups.google.com/d/optout';return true;" class="">https://groups.google.com/d/optout.


--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.
Reply | Threaded
Open this post in threaded view
|

Re: Extreme value distribution

Patrick Kofod Mogensen
Ok, I get it.

On Tuesday, August 25, 2015 at 5:14:29 PM UTC+2, John Myles White wrote:
It's one-off because David has said he doesn't time to maintain it. One-off code can be great code, but even the best code eventually bit rots without active maintenace.

 -- John

On Aug 25, 2015, at 8:07 AM, Patrick Kofod Mogensen <<a href="javascript:" target="_blank" gdf-obfuscated-mailto="-yyY4CrgBwAJ" rel="nofollow" onmousedown="this.href=&#39;javascript:&#39;;return true;" onclick="this.href=&#39;javascript:&#39;;return true;">patrick....@...> wrote:

In what way is it one-off? If it works and is correct, why not PR it to Distributions.jl?

On Friday, August 14, 2015 at 7:19:44 PM UTC+2, David Anthoff wrote:

I coded up a very rough implementation of the Hill (1975) estimator, and then also another method by Huisman et al. (2001). Both methods are short (~10 lines), but they might be the start of a ExtremeValueTheory package.

 

The whole field is not central to what I do, so I would be the wrong person to maintain such a package, but I’d be happy to start it with the two estimators I coded up.

 

Should I just create an unregistered package with name ExtremeValueTheory that has the two algorithms? But people would probably never find it… Or would it make sense to have that package under JuliaStats from the beginning? Or is there some other way to share such one-off implementations with the community?

 

Cheers,

David

 

From: [hidden email] [mailto:julia...@googlegroups.com] On Behalf Of Simon Byrne
Sent: Friday, August 14, 2015 1:24 AM
To: julia-stats <[hidden email]>
Subject: Re: [julia-stats] Extreme value distribution

 

Not at the moment (as far as I know). If you don't want to implement it yourself, you could try RCall.jl.

On Thursday, 13 August 2015 18:09:22 UTC+1, David Anthoff wrote:

But no support for fitting them, right? Thanks, David

 

From: [hidden email] [mailto:julia...@googlegroups.com] On Behalf Of Andreas Noack
Sent: Thursday, August 13, 2015 6:30 AM
To:
[hidden email]
Subject: Re: [julia-stats] Extreme value distribution

 

Hi David,

 

Distributions.jl has the Gumpel, Fréchet and Weibull distributions aka the type I, II and III extreme value distribution.

 

On Wed, Aug 12, 2015 at 5:25 PM, David Anthoff <[hidden email]> wrote:

Hi,

 

is there a package dealing with extreme value distribution stuff? A quick search on <a href="http://pkg.julialang.org/" rel="nofollow" target="_blank" onmousedown="this.href=&#39;http://www.google.com/url?q\75http%3A%2F%2Fpkg.julialang.org%2F\46sa\75D\46sntz\0751\46usg\75AFQjCNGGeneLcTSIu9o4UNv-qZzKc3GIow&#39;;return true;" onclick="this.href=&#39;http://www.google.com/url?q\75http%3A%2F%2Fpkg.julialang.org%2F\46sa\75D\46sntz\0751\46usg\75AFQjCNGGeneLcTSIu9o4UNv-qZzKc3GIow&#39;;return true;">pkg.julialang.org didn’t reveal one. I’m mainly looking for code that estimates the tail index of an empirical data set I have.

 

Thanks,

David

 

--

David Anthoff

University of California, Berkeley

 

<a href="http://www.david-anthoff.com/" rel="nofollow" target="_blank" onmousedown="this.href=&#39;http://www.google.com/url?q\75http%3A%2F%2Fwww.david-anthoff.com%2F\46sa\75D\46sntz\0751\46usg\75AFQjCNHS_ij0xACTsP69xWYtaDWgEwCIgQ&#39;;return true;" onclick="this.href=&#39;http://www.google.com/url?q\75http%3A%2F%2Fwww.david-anthoff.com%2F\46sa\75D\46sntz\0751\46usg\75AFQjCNHS_ij0xACTsP69xWYtaDWgEwCIgQ&#39;;return true;">http://www.david-anthoff.com

 

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to
julia-stats...@googlegroups.com.
For more options, visit
<a href="https://groups.google.com/d/optout" rel="nofollow" target="_blank" onmousedown="this.href=&#39;https://groups.google.com/d/optout&#39;;return true;" onclick="this.href=&#39;https://groups.google.com/d/optout&#39;;return true;">https://groups.google.com/d/optout.

 

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to julia-stats...@googlegroups.com.
For more options, visit <a href="https://groups.google.com/d/optout" rel="nofollow" target="_blank" onmousedown="this.href=&#39;https://groups.google.com/d/optout&#39;;return true;" onclick="this.href=&#39;https://groups.google.com/d/optout&#39;;return true;">https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to julia-stats...@googlegroups.com.
For more options, visit <a href="https://groups.google.com/d/optout" rel="nofollow" target="_blank" onmousedown="this.href=&#39;https://groups.google.com/d/optout&#39;;return true;" onclick="this.href=&#39;https://groups.google.com/d/optout&#39;;return true;">https://groups.google.com/d/optout.


--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to <a href="javascript:" target="_blank" gdf-obfuscated-mailto="-yyY4CrgBwAJ" rel="nofollow" onmousedown="this.href=&#39;javascript:&#39;;return true;" onclick="this.href=&#39;javascript:&#39;;return true;">julia-stats...@googlegroups.com.
For more options, visit <a href="https://groups.google.com/d/optout" target="_blank" rel="nofollow" onmousedown="this.href=&#39;https://groups.google.com/d/optout&#39;;return true;" onclick="this.href=&#39;https://groups.google.com/d/optout&#39;;return true;">https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.
Reply | Threaded
Open this post in threaded view
|

RE: Extreme value distribution

David Anthoff

Everything John said, plus the Hill estimator really doesn’t belong into the Distributions package, it would have to go into an ExtremeValueTheory package (that doesn’t exist). It is one thing to contribute code via a PR to a well maintained package, because even if I didn’t follow up I wouldn’t worry about bit-rot. But starting a new package for one 10 line function that I don’t want to maintain is not a good idea…

 

Having said that, if I post a gist, how would people find it? Maybe we could have a place on the juliastats homepage where people could just link to such one-off code, with a full disclaimer that it is not maintained?

 

Cheers,

David

 

PS: Oh, and I did implement MLE for the pareto distribution in Distributions, so that is there now.

 

From: [hidden email] [mailto:[hidden email]] On Behalf Of Patrick Kofod Mogensen
Sent: Tuesday, August 25, 2015 9:08 AM
To: julia-stats <[hidden email]>
Subject: Re: [julia-stats] Extreme value distribution

 

Ok, I get it.

On Tuesday, August 25, 2015 at 5:14:29 PM UTC+2, John Myles White wrote:

It's one-off because David has said he doesn't time to maintain it. One-off code can be great code, but even the best code eventually bit rots without active maintenace.

 

 -- John

 

On Aug 25, 2015, at 8:07 AM, Patrick Kofod Mogensen <<a href="javascript:" target="_blank">patrick....@...> wrote:

 

In what way is it one-off? If it works and is correct, why not PR it to Distributions.jl?

On Friday, August 14, 2015 at 7:19:44 PM UTC+2, David Anthoff wrote:

I coded up a very rough implementation of the Hill (1975) estimator, and then also another method by Huisman et al. (2001). Both methods are short (~10 lines), but they might be the start of a ExtremeValueTheory package.

 

The whole field is not central to what I do, so I would be the wrong person to maintain such a package, but I’d be happy to start it with the two estimators I coded up.

 

Should I just create an unregistered package with name ExtremeValueTheory that has the two algorithms? But people would probably never find it… Or would it make sense to have that package under JuliaStats from the beginning? Or is there some other way to share such one-off implementations with the community?

 

Cheers,

David

 

From: [hidden email] [[hidden email]] On Behalf Of Simon Byrne
Sent: Friday, August 14, 2015 1:24 AM
To: julia-stats <
[hidden email]>
Subject: Re: [julia-stats] Extreme value distribution

 

Not at the moment (as far as I know). If you don't want to implement it yourself, you could try RCall.jl.

On Thursday, 13 August 2015 18:09:22 UTC+1, David Anthoff wrote:

But no support for fitting them, right? Thanks, David

 

From: [hidden email] [[hidden email]] On Behalf Of Andreas Noack
Sent: Thursday, August 13, 2015 6:30 AM
To:
[hidden email]
Subject: Re: [julia-stats] Extreme value distribution

 

Hi David,

 

Distributions.jl has the Gumpel, Fréchet and Weibull distributions aka the type I, II and III extreme value distribution.

 

On Wed, Aug 12, 2015 at 5:25 PM, David Anthoff <[hidden email]> wrote:

Hi,

 

is there a package dealing with extreme value distribution stuff? A quick search on pkg.julialang.org didn’t reveal one. I’m mainly looking for code that estimates the tail index of an empirical data set I have.

 

Thanks,

David

 

--

David Anthoff

University of California, Berkeley

 

http://www.david-anthoff.com

 

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to
[hidden email].
For more options, visit
https://groups.google.com/d/optout.

 

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.

 

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to <a href="javascript:" target="_blank">julia-stats...@....
For more options, visit https://groups.google.com/d/optout.

 

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.
Reply | Threaded
Open this post in threaded view
|

Re: Extreme value distribution

John Myles White
I would be onboard with a list of unofficial code. I have a lot of packages that could be useful to people, but really shouldn't be listed in METADATA. (Some are there already and really ought to be taken down.)

 -- John

On Aug 25, 2015, at 9:30 AM, David Anthoff <[hidden email]> wrote:

Everything John said, plus the Hill estimator really doesn’t belong into the Distributions package, it would have to go into an ExtremeValueTheory package (that doesn’t exist). It is one thing to contribute code via a PR to a well maintained package, because even if I didn’t follow up I wouldn’t worry about bit-rot. But starting a new package for one 10 line function that I don’t want to maintain is not a good idea…
 
Having said that, if I post a gist, how would people find it? Maybe we could have a place on the juliastats homepage where people could just link to such one-off code, with a full disclaimer that it is not maintained?
 
Cheers,
David
 
PS: Oh, and I did implement MLE for the pareto distribution in Distributions, so that is there now.
 
From: [hidden email] [[hidden email]] On Behalf Of Patrick Kofod Mogensen
Sent: Tuesday, August 25, 2015 9:08 AM
To: julia-stats <[hidden email]>
Subject: Re: [julia-stats] Extreme value distribution
 
Ok, I get it.

On Tuesday, August 25, 2015 at 5:14:29 PM UTC+2, John Myles White wrote:
It's one-off because David has said he doesn't time to maintain it. One-off code can be great code, but even the best code eventually bit rots without active maintenace.
 
 -- John
 
On Aug 25, 2015, at 8:07 AM, Patrick Kofod Mogensen <[hidden email]> wrote:
 
In what way is it one-off? If it works and is correct, why not PR it to Distributions.jl?

On Friday, August 14, 2015 at 7:19:44 PM UTC+2, David Anthoff wrote:
I coded up a very rough implementation of the Hill (1975) estimator, and then also another method by Huisman et al. (2001). Both methods are short (~10 lines), but they might be the start of a ExtremeValueTheory package.
 
The whole field is not central to what I do, so I would be the wrong person to maintain such a package, but I’d be happy to start it with the two estimators I coded up.
 
Should I just create an unregistered package with name ExtremeValueTheory that has the two algorithms? But people would probably never find it… Or would it make sense to have that package under JuliaStats from the beginning? Or is there some other way to share such one-off implementations with the community?
 
Cheers,
David
 
From: [hidden email] [[hidden email]] On Behalf Of Simon Byrne
Sent: Friday, August 14, 2015 1:24 AM
To: julia-stats <
[hidden email]>
Subject: Re: [julia-stats] Extreme value distribution
 
Not at the moment (as far as I know). If you don't want to implement it yourself, you could try RCall.jl.

On Thursday, 13 August 2015 18:09:22 UTC+1, David Anthoff wrote:
But no support for fitting them, right? Thanks, David
 
From: [hidden email] [[hidden email]] On Behalf Of Andreas Noack
Sent: Thursday, August 13, 2015 6:30 AM
To: 
[hidden email]
Subject: Re: [julia-stats] Extreme value distribution
 
Hi David,
 
Distributions.jl has the Gumpel, Fréchet and Weibull distributions aka the type I, II and III extreme value distribution.
 
On Wed, Aug 12, 2015 at 5:25 PM, David Anthoff <[hidden email]> wrote:
Hi,
 
is there a package dealing with extreme value distribution stuff? A quick search on pkg.julialang.org didn’t reveal one. I’m mainly looking for code that estimates the tail index of an empirical data set I have.
 
Thanks,
David
 
--
David Anthoff
University of California, Berkeley
 
 
-- 
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to 
[hidden email].
For more options, visit 
https://groups.google.com/d/optout.
 
-- 
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.
-- 
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.
 

-- 
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.

 
-- 
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.

-- 
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "julia-stats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.