FYI: About Multivariate Temporal Dynamic Analysis

• To: "Statistic Ecology (E-mail)" <STATSOC-L@mhc.mtholyoke.edu>
• Subject: FYI: About Multivariate Temporal Dynamic Analysis
• From: Jay Bai <bai@gpsr.colostate.edu>
• Date: Mon, 20 Apr 1998 18:13:20 -0600
• Organization: GPSR

Hi, Friends,

My commentary on temporal dynamic analysis, 'A critique of matrix solutions for 
ecology' has been published on Bulletin of ESA, Vol.79 No.2., April 1998.

The main points in the commentary are:

1). The samples from a homogeneous community are not linearly independent, 
although they might be statistically independent. Thus, the sample-variable 
matrix should have no inverse.

2). In some cases (mostly in Ecology, Biology, or Agriculture, such as wheat 
and corn), the variables could be mutually independent.

If the above two stands, then no matrix or statistics would work for 
multivariate temporal dynamic analysis. The correct tool would be 
multidimensional space, m-space, and multivariate vectors, m-vectors.

I hope this could be interesting and informative for some friends in StatEco-l. 
Comments and discussions are expected.

Thank you.


T. Jay Bai

• Prev: ESA Stat Section maillist needs a new home
• Next: Third call for judges -- getting desparate
• Index: Mail Index : stat_ecol