Warning

 

Close

Confirm Action

Are you sure you wish to do this?

Confirm Cancel
Member Login
Site Notices
11/22/2017 10:05:29 PM
Posted: 10/21/2004 7:02:52 AM EST
Can you vector me to a good info source on ANOVA? URL/books/papers whatever you got. Say if somebody wants to code this himself, where should I go look? The stuff I found off google are way too general. I am not a statistician but I find myself cornered with a problem that I think needs an ANOVA treatment.

Any experience using the function anovan.m from Matlab toolbox? I keep running out of memory ... my option is either to code one myself (and hopefully I can control the memory management better), or, migrate to SAS .... a real possibility that I dread.

Thanks
oz
Link Posted: 10/21/2004 7:06:28 AM EST
I just do them myself in Excel. Can't help you with Matlab.


I use Modern Mathematical Statictics by Dudewicz and Mishra, or Statistical Quality Control by Montgomery as references.
Link Posted: 10/21/2004 7:15:12 AM EST
I did not know Excel could do anova ..... oh crap, they can do only up to 2 factors .... I need at least 4 factors. Thanks for the info. I'll go look up the books in the library.

oz
Link Posted: 10/21/2004 7:45:27 AM EST
Try:
Data analysis in the Earth Sciences using Matlab by Gerald Middleton. It has a discussion of ANOVA including multiple regressions. I assume this is what you want.

I use Sigmaplot/Sigmastat though.

Link Posted: 10/21/2004 8:00:04 AM EST
Link Posted: 10/21/2004 11:05:20 AM EST

Originally Posted By QCMGR:

What are you trying to do?




I am trying to do data normalization by linear modeling.
Link Posted: 10/21/2004 11:19:53 AM EST
Link Posted: 10/21/2004 11:22:53 AM EST
LOL, Am I the only one who read "Methlab users" at first?
Link Posted: 10/21/2004 11:28:59 AM EST

Originally Posted By MouseGun87:

Originally Posted By QCMGR:

What are you trying to do?




I am trying to do data normalization by linear modeling.



If you are trying to normalize data, why not just do a Box Cox Transformation, or a Johnson Transformation? You can do both of these in Minitab.

Can't help you with Matlab.
Link Posted: 10/21/2004 12:23:50 PM EST
SAS is not that tough to deal with. SPSS is really easy to use if you can get a hole of a copy of SPSS. There may be an add-on Excel module out there that will do more powerful stat functions.

GunLvr
Link Posted: 10/21/2004 5:00:11 PM EST

Originally Posted By PAEBR332:

Originally Posted By MouseGun87:

Originally Posted By QCMGR:

What are you trying to do?




I am trying to do data normalization by linear modeling.



If you are trying to normalize data, why not just do a Box Cox Transformation, or a Johnson Transformation? You can do both of these in Minitab.

Can't help you with Matlab.



I have several factors that may contribute to the variance in the data set. Not sure which one, or if any linear combination of these factors may turn out to be the good discriminating factor. Based on what I have learned so far, you can answer some of these questions using Anova. So there is more to it than just data normalization that I have in my agenda.

I am getting SAS in the next few days. So far, I have Matlab and R at my disposal. Have not tried R yet, but read that people who do similar data processing as I do found out that when Matlab fails to do it due to memory limitation, R does not fare any better. But they seem to agree that SAS will get the job done where the other two had failed.

Thanks to all who replied. Keep it coming.

oz
Top Top