| T O P I C R E V I E W |
| Tryphon |
Posted - 09/14/2016 : 07:51:23 AM
Hello,
I use ANOVA two way repeated measure to study my samples. One of the factors is the time and it has up to 30 levels. I check the interaction box to know the interaction between the sample groups within the time. This represents a big time consuming on my computer and sometimes Origin hangs or crashes.
I am not interested to know the interaction between the time points within a sample group and I wonder if it is possible to skip this calculus which represents a lot of calculations among 30 levels.
Can I set a parameter using Labtalk to skip the calculus of time factor interaction without skipping the calculus of the group factor interaction ?
Here my code :
win -a MyWKS; /* worksheet of my data */
tree twowayGUI;
xop execute:=init classname:=ANOVATwoWayRM iotrgui:=twowayGUI;
twowayGUI.GUI.InputData.Use = 0;
xop execute:=update iotrgui:=twowayGUI;
twowayGUI.GUI.InputData.Factor0$=Col(Time);
twowayGUI.GUI.InputData.Factor1$=Col(Species);
twowayGUI.GUI.InputData.Data$=Col(Data);
twowayGUI.GUI.InputData.Subject$=Col(Identities);
twowayGUI.GUI.Interactions=1;
twowayGUI.GUI.MeansComp.Bonferroni=1;
xop execute:=report iotrgui:=twowayGUI;
page.active=page.nlayers;
Thank you.
Origin Ver. 2015 Pro 64 bits sr2
Operating System: Windows 8.1 64 bits |
| 3 L A T E S T R E P L I E S (Newest First) |
| yuki_wu |
Posted - 09/19/2016 : 03:52:35 AM
Hi,
Thank you for your information.
We will discuss at the meeting later.
Thanks,
Yuki
OriginLab |
| Tryphon |
Posted - 09/19/2016 : 03:30:38 AM
Hello,
I write here my own improvement to cure this issue.
At the beginning of my experiment, there is nothing to observe then I have many time factor levels which contain only 0 values. But I need to put this "null" time levels otherwise my experiment story will not be described entirely and this impacts the statistical results (I tried of course).
My first idea was to replace these 0 by a very small value (0.00001 for example) but the Anova 2 way RM function hanged again.
In fact I had to replace the 0 by the random results given by rnd()/10000 to make the Anova process working! Presently, it cannot support identical values within a level. With this cure, the process never hangs and I get the results faster.
Of course I checked these small values produce the same statistical results as I get in GraphPad Prism which tolerates identical values within a level.
I hope it will help your team.
Thank you. |
| yuki_wu |
Posted - 09/18/2016 : 04:03:03 AM
Hi,
I am sorry that it is not supported for now. I will forward your requirement to our weekly share meeting and see what we can do to improve it.
Sorry for the inconvenience.
Regard,
Yuki
OriginLab
|
