Usually for confidence intervals I just use + 2 std.
So if you have 13,200 observation using 100 iterations, you first find
the std:
stdev =3D std(data, 0, 2)
Then you're left with a standard deviation for each observation using
the 100 iterations.
Then take the mean, plot that, and plot the mean + 2 std.
You can also use the function errorbar after you calculat the standard
deviations.
On Feb 9, 2:50=A0pm, "Kirk" <kwythers.nos...@umn.edu> wrote:
> I am looking for suggestions as to how to approach the problem of plottin=
g =A0continuous confidence intervals of some data in a rather large matrix.=
The matrix dimensions are 13200 x 100. Each column represent an iteration =
of simulated data. Therefor we have 13200 observations (as rows).
>
> I would like to graphically represent all 100 columns by plotting the mea=
n, and then, some kind of continuous confidence interval (say 95%) about th=
e mean line. I am looking at the file 'CONFPLOT' from the file exchange, bu=
t I'm not sure if this is the right approach? Or if it is, how to get it to=
handle all my data columns? Is this a two part step of calculating a mean =
and std across columns, and then calling CONFPLOT, or should I look elsewhe=
re?
