R help - Nov 2011 - Plotting skewed normal distribution with a bar plot

Hi,

I need to create a plot (type = "h")  and then overlay a skewed-normal
curve on this distribution, but I'm not finding a procedure to accomplish
this. I want to use the plot function here in order to control the bin
distributions.

I have explored the sn library and found the dsn function.  dsn uses known
location, scaling and shape parameters associated with a given input vector
of probabilities.  However, how can I calculate the skewed-normal curve if
I don't know these parameters in advance?

Is there another function to calculate the skew-normal, perhaps in a
different package?


I'm working with R 2.13.2 on a windows based machine.

Steve Friedman Ph. D.
Ecologist  / Spatial Statistical Analyst
Everglades and Dry Tortugas National Park
950 N Krome Ave (3rd Floor)
Homestead, Florida 33034

Steve_Friedman at nps.gov
Office (305) 224 - 4282
Fax     (305) 224 - 4147

It seems like you'll need to apply some sort of MLE to estimate the
parameters directly from the data before using dsn() to get the
density. This might help with some of it:
http://help.rmetrics.org/fGarch/html/snorm.html

Michael

On Thu, Nov 3, 2011 at 2:54 PM,  <Steve_Friedman at nps.gov>
wrote:
>
> Hi,
>
> I need to create a plot (type = "h") ?and then overlay a
skewed-normal
> curve on this distribution, but I'm not finding a procedure to
accomplish
> this. I want to use the plot function here in order to control the bin
> distributions.
>
> I have explored the sn library and found the dsn function. ?dsn uses known
> location, scaling and shape parameters associated with a given input vector
> of probabilities. ?However, how can I calculate the skewed-normal curve if
> I don't know these parameters in advance?
>
> Is there another function to calculate the skew-normal, perhaps in a
> different package?
>
>
> I'm working with R 2.13.2 on a windows based machine.
>
> Steve Friedman Ph. D.
> Ecologist ?/ Spatial Statistical Analyst
> Everglades and Dry Tortugas National Park
> 950 N Krome Ave (3rd Floor)
> Homestead, Florida 33034
>
> Steve_Friedman at nps.gov
> Office (305) 224 - 4282
> Fax ? ? (305) 224 - 4147
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>

Steve,

   I don't profess to be an expert in R (by ANY means), but I've used
the sn.mle function in package sn to do something similar. So, for example, if
you have data (y), you can do the following:

y = c(5,4,6,4,5,6,7,7,55,64,23,13,1,4,1,14,1,15,11,12)
sn.mle(y=y)

I just made up this stream of data, so please forgive if it doesn't make
sense.  The function will produce a number of things, including the mles for the
mean, sd, and skew, as well as a histogram with the density overlayed onto it.

Maybe this helps? Hope so.

Greg

-----Original Message-----
>From: "R. Michael Weylandt" <michael.weylandt at gmail.com>
>Sent: Nov 3, 2011 10:39 PM
>To: Steve_Friedman at nps.gov
>Cc: r-help at r-project.org
>Subject: Re: [R] Plotting skewed normal distribution with a bar plot
>
>It seems like you'll need to apply some sort of MLE to estimate the
>parameters directly from the data before using dsn() to get the
>density. This might help with some of it:
>http://help.rmetrics.org/fGarch/html/snorm.html
>
>Michael
>
>On Thu, Nov 3, 2011 at 2:54 PM,  <Steve_Friedman at nps.gov> wrote:
>>
>> Hi,
>>
>> I need to create a plot (type = "h") ?and then overlay a
skewed-normal
>> curve on this distribution, but I'm not finding a procedure to
accomplish
>> this. I want to use the plot function here in order to control the bin
>> distributions.
>>
>> I have explored the sn library and found the dsn function. ?dsn uses
known
>> location, scaling and shape parameters associated with a given input
vector
>> of probabilities. ?However, how can I calculate the skewed-normal curve
if
>> I don't know these parameters in advance?
>>
>> Is there another function to calculate the skew-normal, perhaps in a
>> different package?
>>
>>
>> I'm working with R 2.13.2 on a windows based machine.
>>
>> Steve Friedman Ph. D.
>> Ecologist ?/ Spatial Statistical Analyst
>> Everglades and Dry Tortugas National Park
>> 950 N Krome Ave (3rd Floor)
>> Homestead, Florida 33034
>>
>> Steve_Friedman at nps.gov
>> Office (305) 224 - 4282
>> Fax ? ? (305) 224 - 4147
>>
>> ______________________________________________
>> R-help at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-help
>> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
>> and provide commented, minimal, self-contained, reproducible code.
>>
>
>______________________________________________
>R-help at r-project.org mailing list
>https://stat.ethz.ch/mailman/listinfo/r-help
>PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
>and provide commented, minimal, self-contained, reproducible code.

You might want to check again: I'm running fGarch on 2.13.2, Mac OSX 10.5.8.

Michael

On Fri, Nov 4, 2011 at 9:36 AM, Steve Friedman <skfglades at gmail.com>
wrote:
> Hi Michael,
>
> Thanks for pointing me fGarch.? I actually started there, but it is not yet
> available for 2.13.2 so I went directly to the (sn-package).
>
> I've briefly explored your suggestion and think it will work.
>
> Thanks
> Steve
>
> On Nov 3, 2011 10:41 PM, "R. Michael Weylandt"
<michael.weylandt at gmail.com>
> wrote:
>>
>> It seems like you'll need to apply some sort of MLE to estimate the
>> parameters directly from the data before using dsn() to get the
>> density. This might help with some of it:
>> http://help.rmetrics.org/fGarch/html/snorm.html
>>
>> Michael
>>
>> On Thu, Nov 3, 2011 at 2:54 PM, ?<Steve_Friedman at nps.gov>
wrote:
>> >
>> > Hi,
>> >
>> > I need to create a plot (type = "h") ?and then overlay a
skewed-normal
>> > curve on this distribution, but I'm not finding a procedure to
>> > accomplish
>> > this. I want to use the plot function here in order to control the
bin
>> > distributions.
>> >
>> > I have explored the sn library and found the dsn function. ?dsn
uses
>> > known
>> > location, scaling and shape parameters associated with a given
input
>> > vector
>> > of probabilities. ?However, how can I calculate the skewed-normal
curve
>> > if
>> > I don't know these parameters in advance?
>> >
>> > Is there another function to calculate the skew-normal, perhaps in
a
>> > different package?
>> >
>> >
>> > I'm working with R 2.13.2 on a windows based machine.
>> >
>> > Steve Friedman Ph. D.
>> > Ecologist ?/ Spatial Statistical Analyst
>> > Everglades and Dry Tortugas National Park
>> > 950 N Krome Ave (3rd Floor)
>> > Homestead, Florida 33034
>> >
>> > Steve_Friedman at nps.gov
>> > Office (305) 224 - 4282
>> > Fax ? ? (305) 224 - 4147
>> >
>> > ______________________________________________
>> > R-help at r-project.org mailing list
>> > https://stat.ethz.ch/mailman/listinfo/r-help
>> > PLEASE do read the posting guide
>> > http://www.R-project.org/posting-guide.html
>> > and provide commented, minimal, self-contained, reproducible code.
>> >
>>
>> ______________________________________________
>> R-help at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-help
>> PLEASE do read the posting guide
>> http://www.R-project.org/posting-guide.html
>> and provide commented, minimal, self-contained, reproducible code.
>