| baseini {unknown} | R Documentation |
This function can be used to derive initial values for the
baseline function which are required by two other functions:
puldet and
pulest.
baseini(x, y, data, puloc, method=c("shift", "select"),
smooth=c("spline", "loess"), control=list(shift=-1, spline=list(nb=~x,
rk=cubic(x), spar="v", limnla=c(-1.5, 0)), loess=list(span=0.75),
select=list(npl=2, npr=3)))
x |
a vector of observation time points. |
y |
a vector of hormone concentrations. |
data |
a data frame containing the variables occurring in the x and y arguments. If this option is not specified, the variables should be on the search list. Missing values are not allowed. |
puloc |
a vector of pulse locations. |
method |
a character string. If method="shift", a smooth curve will be fit to all observations using a smoothing method specified by the smooth argument, and then the whole curve will be shifted downward; if method="select", a smooth curve will be fit to the selected observations without those around pulse locations. |
smooth |
a character string specifying which smoothing method
will be used:
smooth="spline" for smoothing spline and smooth="loess" for loess. |
control |
a list of control parameters for different method
and smooth options. Names of the list are shift, spline,
select, and loess. When method="shift", shift
specifies the amount of shifting downward with a nonnegative value; the
default shift=-1 asks the program to estimate the amount of
shift. When
method="select", the select list in the control argument
decides the number of observation points to be eliminated on the
left (npl) and on the right (npr) of each pulse location.
When smooth="spline", as in pcp, the spline
list in the control argument specifies the null space, reproducing kernel, and
method and range for estimating the smoothing parameter.
When smooth="loess", the loess list in the control argument
specifies the span for the loess function. |
When smooth="spline", the ssr function
in the assist package is used to fit the spline function.
See cran.us.r-project.org/doc/packages/assist.pdf
for more information about nb, rk, spar and
limnla. The appropriate choices of npl and npr depend
on the sampling rate as well as prior knowledge about the infusion and
decay rates.
Estimated baseline function evaluated at x.
Yu-Chieh Yang, Anna Liu, Yuedong Wang
Wang, Y. and Ke, C. (2002), ASSIST: A Suite of S-plus functions Implementing Spline smoothing Techniques. Available at cran.us.r-project.org/src/contrib/PACKAGES.html. Manual for the ASSIST package is available at www.pstat.ucsb.edu/faculty/yuedong/software.
Yang, Y. (2002), Detecting Change Points and Hormone Pulses Using Partial Spline Models, Ph.D. Thesis, University of California-Santa Barbara, Dept. of Statistics and Applied Probability.
loess, ssr
bl1 <- baseini(time, conc, data=acth, puloc=pl1, method="shift", smooth="spline") bl2 <- baseini(time, conc, data=acth, puloc=pl1, method="select", smooth="spline") bl3 <- baseini(time, conc, data=acth, puloc=pl1, method="shift", smooth="loess") bl4 <- baseini(time, conc, data=acth, puloc=pl1, method="select", smooth="loess")