home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!europa.asd.contel.com!paladin.american.edu!auvm!UNLVM.BITNET!BIOM030
- Message-ID: <STAT-L%93012211013435@VM1.MCGILL.CA>
- Newsgroups: bit.listserv.stat-l
- Date: Fri, 22 Jan 1993 09:49:59 CST
- Sender: STATISTICAL CONSULTING <STAT-L@MCGILL1.BITNET>
- From: April Milliken <BIOM030@UNLVM.BITNET>
- Subject: Multiple Mulitvariate Analyses???
- Lines: 28
-
- I am still in the "playing" stage with a set of data, but I am curious if
- anyone has experience with what combining several multivariate analyses?
- I have principle component analysis results for 12 different time periods,
- on the same variables, same animals. I am really intersted in the OVERALL
- effect of treatment, but time is also a factor. How can I combine my results?
- Or, is there a better way? So far, PCA is similar for components 1 and 2,
- but from there it starts getting quite variable. Somewhere in the range of
- 75% to 90% of total variation is accounted for by PC 3 (over all time periods).
- So, 3 PC's may be enough. AND, the biology makes sense (for the most part)
- for each time period. Now, is there a way to model what is going on over
- time, across the PCA's? Am I asking too much? I need some ideas!
-
- If you have even a small suggestion, please say so. I am open to any new
- direction.
-
- Thanks,
- April
-
- ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
-
- April Milliken
-
- Department of Biometry
- University of Nebraska-Lincoln
-
- BIOM030@UNLVM.UNL.EDU
-
- "I may be in Nebraska, but my heart is still at Kansas State University!"
-