Produces a chart of the original time series along with the fitted values. Optionally, predicted values (and their confidence bounds) can also be plotted.
# S3 method for HoltWinters
plot(x, predicted.values = NA, intervals = TRUE,
separator = TRUE, col = 1, col.predicted = 2,
col.intervals = 4, col.separator = 1, lty = 1,
lty.predicted = 1, lty.intervals = 1, lty.separator = 3,
ylab = "Observed / Fitted",
main = "Holt-Winters filtering",
ylim = NULL, …)Object of class "HoltWinters"
Predicted values as returned by predict.HoltWinters
If TRUE, the prediction intervals are plotted (default).
If TRUE, a separating line between fitted and predicted values is plotted (default).
Color/line type of original data (default: black solid).
Color/line type of fitted and predicted values (default: red solid).
Color/line type of prediction intervals (default: blue solid).
Color/line type of observed/predicted values separator (default: black dashed).
Label of the y-axis.
Main title.
Limits of the y-axis. If NULL, the range is chosen
such that the plot contains the original series, the fitted values,
and the predicted values if any.
Other graphics parameters.
C. C. Holt (1957) Forecasting trends and seasonals by exponentially weighted moving averages, ONR Research Memorandum, Carnegie Institute of Technology 52.
P. R. Winters (1960). Forecasting sales by exponentially weighted moving averages. Management Science, 6, 324--342. 10.1287/mnsc.6.3.324.