| Author |
 Topic  |
|
|
Enantio
Germany
7 Posts |
 Posted - 07/02/2011 : 2:27:46 PM
|
Origin Ver. and Service Release: 8.5
Operating System: WinXP
I am doing statistics on different particle sizes and was wondering how I could make two normal curves if the data appears to consist of a bimodal distribution.
I thought about splitting the data and making two layers with the smaller and the bigger particles respectively but I am not so pleased with this idea. So is there a more decent way to cope with this issue.
Thanks in advance. |
|
|
Hideo Fujii
USA
1582 Posts |
Posted - 07/05/2011 : 2:07:34 PM
|
Hi Enantio,
Not sure if the following works for you, but you can consider the sample below to make a two-peak Gaussian curve from a dataset for frequencies.
1) Make a histogram from the dataset. (Graph1)
2) Double-click the histogram to show Plot Details dialog. Choose "Data" tab, and press"Go to Bin Worksheet" button (to make the second sheet in the workbook).
3) Highlight "Count" column (against Bin Centers), and plot it as a scatter plot. (Graph2)
4) For the scatter plot, choose "Analysis: Peaks and Baseline: Multiple Peak Fit" menu. Select the two center positions on the graph (as the initial estimated locations), and get the result. In the result worksheet, you can get parameters of each curve like the centers and widths.
Hope the above procedure is helpful.
--Hideo Fujii
OriginLab |
Edited by - Hideo Fujii on 07/05/2011 2:24:54 PM |
 |
|
|
Enantio
Germany
7 Posts |
Posted - 07/06/2011 : 05:07:56 AM
|
Hi Hideo,
thank you for this very nice solution to my problem. It works quite well. 
I just was wondering, about the "Go to Bin Worksheet" button, why does it use different bin sizes than in the histogram (for example mine are 2.5, but in the worksheet the bin size is 5)? It's not such a big problem, I could circumvent it with a manual frequency count on the data column and then proceed as you suggested.
|
|
|
|
Hideo Fujii
USA
1582 Posts |
Posted - 07/07/2011 : 3:55:53 PM
|
Hi Enantio,
> about the "Go to Bin Worksheet" button, why does it use
> different bin sizes than in the histogram (for example mine
> are 2.5, but in the worksheet the bin size is 5)?
I have tried in my Origin 8.5.1 SR0, but I couldn't reproduce the problem you have pointed out.
Maybe you can send me (tech@originlab.com with ATTN:Hideo) the Origin file so that
I can try and look at it.
--Hideo Fujii
OriginLab |
|
|
|
bibrahim
USA
1 Posts |
Posted - 02/22/2019 : 6:57:09 PM
|
Hello,
After I did the fit, there is an ANOVA test. What does this mean? Does it test the significance of the fitting?
Thanks,
Baher |
|
|
| |
Topic |
|
|
|
Bimodal Distribution