[gmx-users] Reg: Standard error

Mark Abraham Mark.Abraham at anu.edu.au
Wed Dec 1 16:47:40 CET 2010


On 1/12/2010 8:25 PM, vinothkumar mohanakrishnan wrote:
> Dear Mark
>
> Thank you for your explaination but Its not clear to me what you are 
> trying to say. can you explain it bit more clearly. what should i need 
> to do to get the standard error? any help is highly appreciated.

Like I said, you need to make independent measurements of a quantity to 
form an estimate of the mean, and thus an estimate of the standard 
deviation among such estimates (i.e. standard error of the mean). This 
is pretty basic scientific statistics. Your MD steps are all correlated 
because of how they were generated, so you need to look at the 
autocorrelation time (Google!) of your observable to get a starting idea 
of how many actual independent measurements your data represents.

Mark

>
> Regards
> Vinoth
>
> On Wed, Dec 1, 2010 at 2:39 PM, Mark Abraham <Mark.Abraham at anu.edu.au 
> <mailto:Mark.Abraham at anu.edu.au>> wrote:
>
>     On 1/12/2010 7:59 PM, vinothkumar mohanakrishnan wrote:
>
>         Hi Gromacians
>
>         I am planning to compute the standard error for the
>         interfacial surface tension of water-dichloro ethane. i just
>         want to know is my procedure right. below is the result what i
>         get from my g_energy command.
>
>         Statistics over 500001 steps [ 0.0000 thru 1000.0001 ps ], 1
>         data sets
>         All averages are exact over 500001 steps
>
>         Energy                      Average       RMSD     Fluct.    
>          Drift  Tot-Drift
>         -------------------------------------------------------------------------------
>         #Surf*SurfTen               692.904    4579.46    4579.16
>          -0.181601   -181.602
>
>         my question is to calculate the standard error should i need
>         to divide the RMSD(4579.46) by the squareroot of 500001 steps?
>         any help is highly appreciated.
>
>
>     Calculating the standard error of the mean in this way requires
>     that the observations are independent
>     (http://en.wikipedia.org/wiki/Standard_error_%28statistics%29).
>     Adjacent MD frames are not. You would need to know something about
>     your correlation times to know how long you have to wait before
>     effective independence.
>
>     Mark
>     -- 
>     gmx-users mailing list gmx-users at gromacs.org
>     <mailto:gmx-users at gromacs.org>
>     http://lists.gromacs.org/mailman/listinfo/gmx-users
>     Please search the archive at
>     http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
>     Please don't post (un)subscribe requests to the list. Use the www
>     interface or send it to gmx-users-request at gromacs.org
>     <mailto:gmx-users-request at gromacs.org>.
>     Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20101202/7693c675/attachment.html>


More information about the gromacs.org_gmx-users mailing list