Skip to content

data(lh) time serie parameters

2 messages · Gabor Grothendieck, Jean lobry

Original
Gabor Grothendieck
# 
On Sun, Mar 30, 2008 at 5:52 AM, Jean lobry
<lobry at biomserv.univ-lyon1.fr> wrote:
It seems they are using 10 minutes as the unit of measurement.
If you wish to change it to hours you might want to use this instead:

   lh.hr <- ts(lh, start = 0, frequency = 6)

so that

   cycle(lh.hr)

starts out at 1.
1 day later
Jean lobry
# 
Dear Gabor,

yes, it is very likely that the original dataset was with a 10 minutes
unit of measurement because plot(lh[1:44]) yields the same curve as
figure 1b in:

@article{DigglePJ1991,
     title = {Nonparametric comparison of cumulative periodograms},
     author = {Diggle, P.J. and Fisher, N.I.},
     journal = {Applied Statistics},
     volume = {40},
     pages = {423-434},
     year = {1991}
}

whose legend reads:

QUOTE
Fig. 1. Four time series of LH concentrations in blood samples: each
series consists of $n = 44$ values derived from blood samples taken at
10-min intervals from a healthy adult female (data collected by Dr A.
Murdoch): a) early follicular phase, first cycle; (b) early follicular
phase, second cycle; (c) late follicular phase, first cycle; (d) late
follicular phase, second cycle
UNQUOTE

The complete multiple time series with 48 values as in data(lh) was
deposited in statlib by P.J. Diggle himself:
http://lib.stat.cmu.edu/datasets/diggle
(note that the first one has 4 trailling 0 that look like NA which
may explain why there were only 44 values in the 1991 paper).

What I found confusing is that

plot(lh)

ends at 48 when the doc says that lh is a "regular time series".
That would be OK for me if lh was a simple vector, but lh is a
time series so I would expect it to have different tsp() parameters
to fit the doc.

Best,

Jean