Maximum Entropy Modelling Tutorial

[ud...@hotmail.com]

Dear MaxEnt Users,

Colleagues and I have put together this Microsoft Excel-based Tutorial
on Maximum-Entropy Species-Distribution modelling, based on Steven
Phillip's and colleagues papers.

http://research.microsoft.com/en-us/downloads/8d287b15-b463-44e8-a35f-95ffbbff8459/

It was initially done to just get our head around MaxEnt; the
assumptions, maths and parameters. However we've put it online to ease
the learning curve for others in future (hopefully). If you spot any
mistakes, please let me know (I am Matthew Smith at Microsoft Research
in Cambridge, UK).

Best wishes,
M

[João Paulo Soares de Cortes]

I hadn´t  had sucess in openning the file mate. any suggest?
 Cheers !
 
> Date: Mon, 25 Oct 2010 01:05:35 -0700
> Subject: Maximum Entropy Modelling Tutorial
> From: ud...@hotmail.com
> To: max...@googlegroups.com

> --
> You received this message because you are subscribed to the Google Groups "Maxent" group.
> To post to this group, send email to max...@googlegroups.com.
> To unsubscribe from this group, send email to maxent+un...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/maxent?hl=en.
>

[ud...@hotmail.com]

Hi Joao,

Ah - two things. i) It's been compressed as a .zip file so you'll need
to unzip it using Winzip 7-zip or something like that. ii) Once
unzipped, it becomes a Microsoft Excel file - but in the new format.
With .xlsx on the end. You need Office 2007 or later I think to view
those (older versions were .xls). Let me know if you want it in the
older format (or get someone to save it for you in the older format) -
Email me if you want the file.

Best,
M

On Oct 26, 4:49 am, João Paulo Soares de Cortes

<tweedled...@hotmail.com> wrote:
> I hadn´t  had sucess in openning the file mate. any suggest?
>  Cheers !
>
>
>
> > Date: Mon, 25 Oct 2010 01:05:35 -0700
> > Subject: Maximum Entropy Modelling Tutorial
> > From: ud...@hotmail.com
> > To: max...@googlegroups.com
>
> > Dear MaxEnt Users,
>
> > Colleagues and I have put together this Microsoft Excel-based Tutorial
> > on Maximum-Entropy Species-Distribution modelling, based on Steven
> > Phillip's and colleagues papers.
>

> >http://research.microsoft.com/en-us/downloads/8d287b15-b463-44e8-a35f...

>
> > It was initially done to just get our head around MaxEnt; the
> > assumptions, maths and parameters. However we've put it online to ease
> > the learning curve for others in future (hopefully). If you spot any
> > mistakes, please let me know (I am Matthew Smith at Microsoft Research
> > in Cambridge, UK).
>
> > Best wishes,
> > M
>
> > --
> > You received this message because you are subscribed to the Google Groups "Maxent" group.
> > To post to this group, send email to max...@googlegroups.com.
> > To unsubscribe from this group, send email to maxent+un...@googlegroups.com.

> > For more options, visit this group athttp://groups.google.com/group/maxent?hl=en.- Hide quoted text -
>
> - Show quoted text -

[John B]

Check out the MS Office Compatibility Pack, which should allow you to
open .xlsx in pre-2007 versions of Office: http://support.microsoft.com/kb/924074

John

> > > For more options, visit this group athttp://groups.google.com/group/maxent?hl=en.-Hide quoted text -

[Babak Naimi]

Dear Matthew,

I have a question:
In sheet '5. Fit Model', in the cell in which the model likelihood is
calculated 'G25', I can't understand the last part of the formula:
=SUMIF('4. Training calculations'!$E$27:$E$126,1,'4. Training
calculations'!$S$27:$S$126)/Table8[[#Totals],[Species present?]]-LN(1/
RawData[Site ID number])
I mean, 'RawData[Site ID number]'. the first part averages the log
probabilities at presence sites, but second part, i have no idea what
that is for... in this cell 'RawData[Site ID number]' retuns 13, which
is the value of 'Site ID number' field in RawData table in row of 25
(the row that this cell is located)!

I appreciate if you provide me an explanation.

By the way, this toturial is great, thanks a lot.

Best wishes,
Babak

On Oct 25, 10:05 am, "ud...@hotmail.com" <ud...@hotmail.com> wrote:
> Dear MaxEnt Users,
>
> Colleagues and I have put together this Microsoft Excel-based Tutorial
> on Maximum-Entropy Species-Distribution modelling, based on Steven
> Phillip's and colleagues papers.
>

> http://research.microsoft.com/en-us/downloads/8d287b15-b463-44e8-a35f...

[ud...@hotmail.com]

Dear Babak,
Thank you for spotting this. This additional term has a mistake. I put
it in when experimenting to figure out what the "regularized training
gain" meant. It was intended to be coded up as LN(1/COUNT(RawData[Site
ID number])) - to control for the likelihood associated with choosing
a cell correctly at random. The difference is the inclusion of the
COUNT function. I was trying to get the same result to what MaxEnt
gave in it's "gain" metric (with a difference in sign). For now I
simply recommend removing this term (-LN(1/RawData[Site ID number]))
to make it identical (I think) to the cited papers - you should find
that the fit is the same anyway because it was effectively removign a
constant from the estimated likelihood.
Thanks a lot for spotting this and please feel free to highlight any
other bugs. I'll update the download soon.
Best,
M

> > M- Hide quoted text -

[Steve Garske]

Hello Matthew,

Looks like there's a minor notational error in worksheet 4 (training
calculations). The column heading in table 4, "exp(sum(l.f(x)))" should read
"exp(sum(l.f(x))minus const)". Still working through the tutorial. This
tutorial is very helpful - thanks!

Steve
GLIFWC
Odanah, WI USA

On Oct 25, 10:05 am, "ud...@hotmail.com" <ud...@hotmail.com> wrote:

> Dear MaxEnt Users,
>
> Colleagues and I have put together this Microsoft Excel-based Tutorial
> on Maximum-Entropy Species-Distribution modelling, based on Steven
> Phillip's and colleagues papers.
>

> http://research.microsoft.com/en-us/downloads/8d287b15-b463-44e8-a35f...

>
> It was initially done to just get our head around MaxEnt; the
> assumptions, maths and parameters. However we've put it online to ease
> the learning curve for others in future (hopefully). If you spot any
> mistakes, please let me know (I am Matthew Smith at Microsoft Research
> in Cambridge, UK).
>
> Best wishes,
> M

--

[ud...@hotmail.com]

Thanks Steve,
You are of course right. This operation is to ensure that the sum of
the weighted feature values is always less than or equal to zero -
otherwise you might get "suitability" values of greater than 1.
Thanks a lot for your attention to detail and glad you're finding it
useful. More of the same welcome. I'll update the download soon.
Best,
M

> For more options, visit this group athttp://groups.google.com/group/maxent?hl=en.- Hide quoted text -

[Fateme Falah]

hello all
how can I download maxnet software?
I appreciate any help
falahf...@gmail.com

در دوشنبه 25 اکتبر 2010، ساعت 11:35:35 (UTC+3:30)، Matthew Smith نوشته: