| Author |
 Topic  |
|
|
GaussianFit
38 Posts |
 Posted - 09/06/2012 : 02:26:27 AM
|
Origin Ver. and Service Release (Select Help-->About Origin): 8.0 SR6
Operating System: Windows XP
Calling for help on the following problem:
I've got multiple curves plotted in the same X-Y graph. Each curve is taken from X and Y columns in distinctive workbooks. Therefore each curve may have its own X data set with different length and values. A statistics on all the curves needs to be done to calculate an "average curve" with Y values as the mean values of all the Y values and error bars as the standard deviations. Note that some Y values may need to be interpolated first. Preferably the user could select the X-axis range for such statistics (by mouse drag or type in a range), or just use the maximal X-axis span that covers all the curves.
Thanks for any suggestions. |
|
|
Hideo Fujii
USA
1582 Posts |
Posted - 09/06/2012 : 3:15:08 PM
|
Hi GaussianFit,
Not sure whether you want to create an "average" curve among multiple curves taking
account of both X's and Y's; or want to calculate the meas and SD of Y's on each plot.
For the later case, you have a chance to use OriginPro's version 8.5.1, or later,
you can try the "Cluster Gadget", which can interactively move the ROI box to set the
XY range flexibly. It reports the descriptive statistics of each dataset, and of
all data points together.
For the first case, as you indicated, you first need to get interpolated curves in order to
let them have the common X positions.
--Hideo Fujii
OriginLab
|
 |
|
|
GaussianFit
38 Posts |
Posted - 09/06/2012 : 4:24:35 PM
|
Below is an simplfied example:
Three curves are plotted in a graph
Curve 1: Y1=X1^0.5+ran(); X1=[0.2,0.4,0.6,0.8,1.0]
Curve 2: Y2=0.9*X2^0.51+ran(); X2=[0.1,0.3,0.5,0.7,0.9]
Curve 3: Y3=0.8*X3^0.52+ran(); X3=[0.15,0.35,0.55,0.75,0.95]
ran() was used to simulate noise in the actual data.
X1, X2, X3 are made different as well.
Now I need to calulate Y as mean value of Y1, Y2, Y3 at each
interpolated X point for a new curve, and calculate SD of Y
reflecting the spreading of Y1, Y2, Y3 values. The range of X can be
user selected.
It's great to hear Origin 8.5 has similar function, but I don't have a chance to try Origin 8.5 though. Can something be figured out in
Origin 8.0?
quote:
Originally posted by Hideo Fujii
Hi GaussianFit,
Not sure whether you want to create an "average" curve among multiple curves taking
account of both X's and Y's; or want to calculate the meas and SD of Y's on each plot.
For the later case, you have a chance to use OriginPro's version 8.5.1, or later,
you can try the "Cluster Gadget", which can interactively move the ROI box to set the
XY range flexibly. It reports the descriptive statistics of each dataset, and of
all data points together.
For the first case, as you indicated, you first need to get interpolated curves in order to
let them have the common X positions.
--Hideo Fujii
OriginLab
|
|
|
| |
Topic |
|
|
|
Statistics on multiple curves in a graph