On 6 Feb 2003 14:42:12 -0800, [EMAIL PROTECTED] (Laurent Therond) wrote: > [Question] > How to apply the rules of correlation and linear regression to a data > set constituting a discrete time series? > > In other words, given a data set such as the following: > > 2002/1/1, 4500.63 > 2002/1/2, 10255.36 > 2002/1/8, 6530.63 > 2002/1/9, 5230.36 > ...
You can convert the dates to a sequence of day-numbers. It does not matter, statistically speaking, whether the intervals between two dates are in days, hours, years, or (even) z-scores of the date-variable. I don't think of an excuse for using z-scores, but *they* would contain the same information. ANOVAs are indifferent to additive and multiplicative transformation, and that includes Ordinary Least Squares regression It appears that if you use "days" or "weeks" for the ones above, it will be easiest to read or back-translate the eventual effect size. [ snip, some] > If I have 2 time series resulting from the observation of a same > phenomenon, how should I build their respective regression line so > that I may compare them? (granted that scales of measurements could > be different and that dates at which those measurements are taken > could be different) [ snip, rest] If, as you say, your interest is in description rather than in testing, then you might settle for giving two curves and some semblance of their errors-of-prediction. I have discovered on previous problems that the proper advice is often highly dependent on a couple of things that you don't mention -- Ns, and R-squared. Also, what is your purpose? What I just said about curves is more reasonable if the auto-correlations are "low" rather than "high". Are the first-order correlations on the order of 0.01? 0.20? 0.50? 0.90? 0.99? Are the Ns on the order of 10? 100? 1000? -- Rich Ulrich, [EMAIL PROTECTED] http://www.pitt.edu/~wpilib/index.html . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
