cobs (version 1.3-9)

DublinWind: Daily Wind Speeds in Dublin

Description

The DublinWind data frame is basically the time series of daily average wind speeds from 1961 to 1978, measured in Dublin, Ireland. These are 6574 observations (18 full years among which four leap years).

Usage

data(DublinWind)

Format

This data frame contains the following columns:

speed

numeric vector of average daily wind speed in knots

day

an integer vector giving the day number of the year, i.e., one of 1:366.

Details

The periodic pattern along the 18 years measured and the autocorrelation are to be taken into account for analysis, see the references. This is Example 3 of the COBS paper.

References

Haslett, J. and Raftery, A. (1989) Space-Time Modelling with Long-Memory Dependence: Assessing Ireland's Wind Power Resource (with Discussion: 22-50). Applied Statistics 38, 1--50. tools:::Rd_expr_doi("10.2307/2347679")

COBS: Qualitatively Constrained Smoothing via Linear Programming; Computational Statistics 14, 315--337.

He, X. and Ng, P. (1999) COBS: Qualitatively Constrained Smoothing via Linear Programming; Computational Statistics 14, 315--337.

Examples

data(DublinWind)
str(DublinWind)
plot(speed ~ day, data = DublinWind)# not so nice; want time scale

## transform 'day' to correct "Date" object; and then plot
Dday <- seq(from = as.Date("1961-01-01"), by = 1,
            length = nrow(DublinWind))
plot(speed ~ Dday, data = DublinWind, type = "l",
     main = paste("DublinWind speed daily data, n=",
                  nrow(DublinWind)))

##--- ~ He & Ng "Example 3" %% much more is in ../tests/wind.R
co.o50 <-
 with(DublinWind, ## use nknots > (default) 6 :
  cobs(day, speed, knots.add = TRUE, constraint= "periodic", nknots = 10,
       tau = .5, method = "uniform"))
summary(co.o50)
lines(Dday, fitted(co.o50), col=2, lwd = 2)

## the periodic "smooth" function - in one period
plot(predict(co.o50), type = "o", cex = 0.2, col=2,
     xlab = "day", ylim = c(0,20))
points(speed ~ day, data = DublinWind, pch = ".")