How to skip/ignore NaN's when calculating means
in [Matlab]
|
From: Kirk on 4 May 2010 12:45 I am running multiple iterations of a simulation process and then outputting means from all iterations with a simple call like this: % Calculate means from Monte Carlo runs tmax_mean=mean(tmax,2); Every so often... one time in a thousand or so, the random combination of parameters results in NaN output. Which of course, results in NaN means. I am looking for a MATLABish approach to calculate the mean of the variable tmax, but skip the NaNs. Thanks in advance
From: Sean on 4 May 2010 13:02 "Kirk" <kwythers.nospam(a)umn.edu> wrote in message <hrpiuv$1nq$1(a)fred.mathworks.com>... > I am running multiple iterations of a simulation process and then outputting means from all iterations with a simple call like this: > > % Calculate means from Monte Carlo runs > tmax_mean=mean(tmax,2); > > Every so often... one time in a thousand or so, the random combination of parameters results in NaN output. Which of course, results in NaN means. > > I am looking for a MATLABish approach to calculate the mean of the variable tmax, but skip the NaNs. > > Thanks in advance >>help nanmean
From: us on 4 May 2010 13:45 "Kirk" <kwythers.nospam(a)umn.edu> wrote in message <hrpiuv$1nq$1(a)fred.mathworks.com>... > I am running multiple iterations of a simulation process and then outputting means from all iterations with a simple call like this: > > % Calculate means from Monte Carlo runs > tmax_mean=mean(tmax,2); > > Every so often... one time in a thousand or so, the random combination of parameters results in NaN output. Which of course, results in NaN means. > > I am looking for a MATLABish approach to calculate the mean of the variable tmax, but skip the NaNs. > > Thanks in advance a hint: - if(f) you own the stats tbx help nanmean; % <- and siblings... us
From: Kirk on 4 May 2010 13:51 > >>help nanmean Wow. That's nice!
|
Pages: 1 Prev: APPCOEF IN WAVELET TOOL BOX Next: Quant Developer Needed; Top Hedge Fund; London |