[R-sig-Epi] seeking helpful suggestions from R-users, please

Bonn K. Shalf bookshlf at hotmail.com
Fri Apr 6 20:01:35 CEST 2007


Hi all,

New to R and this list and hoping to get some helpful suggestions from 
R-users.

Wondered if users have experience with time series analysis methods for 
irregular (in time) data?  I have access to normalized data for seizure 
studies of patients from four groups and want to examine correlations (or 
degrees of less than 100% correlation) between these irregular, asynchronous 
time series.  The hope is to see what subset of signals might possibly be 
triggers to other signals.  Possibly R-users may have experience with 
appropriate methods?  Maybe someone could suggest a method I might use which 
is more appropriate than a Wald-Granger-type causality test, which doesn't 
require interpolation to a regular-in-time time structure, and examine raw, 
irregular data?

I'm in an awkward situation (advisor has been denied tenure, leaving the 
department on quite difficult terms in May, and me without funding - TA or 
RA - since Dec. 2006), and I hoped it wouldn't come to asking for help from 
scientists outside my department.  Was hoping to make this masters a 
springboard to a PhD in biostatistics or epidemiological science but my 
current focus is just to get my degree.

Thank you for considering this posting, and my apologies if some of you find 
this posting inappropriate, along the lines of, "Help me Obi Wan Kenobi, 
you're my only hope..."

B. Shalf



More information about the R-sig-Epi mailing list