1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
|
\name{apply}
\alias{apply}
\alias{apply,timeSeries-method}
\alias{fapply}
\alias{applySeries}
\alias{rollDailySeries}
\title{Apply functions over time windows}
\description{
Applies a function to a \code{"timeSeries"} object over regular or
irregular time windows, possibly overlapping.
}
\usage{
\S4method{apply}{timeSeries}(X, MARGIN, FUN, \dots, simplify = TRUE)
fapply(x, from, to, FUN, \dots)
applySeries(x, from = NULL, to = NULL, by = c("monthly", "quarterly"),
FUN = colMeans, units = NULL, format = x@format,
zone = x@FinCenter, FinCenter = x@FinCenter,
recordIDs = data.frame(), title = x@title,
documentation = x@documentation, \dots)
rollDailySeries(x, period = "7d", FUN, \dots)
}
\arguments{
\item{x,X}{
an object of class \code{timeSeries}.
}
\item{MARGIN}{
a vector giving the subscripts which the function will be
applied over, see base R's \code{\link[base]{apply}}.
}
\item{FUN}{
the function to be applied. For the function \code{applySeries} the
default setting is \code{FUN = colMeans}.
}
\item{simplify}{
simplify the result?
}
\item{from, to}{
starting date and end date as \code{"timeDate"} objects. Note,
\code{to} must be time ordered after \code{from}. If \code{from}
and \code{to} are missing in function \code{fapply} they are set by
default to \code{from=start(x)}, and \code{to=end(x)}.
}
\item{by}{
a character value either \code{"monthly"} or \code{"quarterly"} used
in the function \code{applySeries}. The default value is
\code{"monthly"}. Only operative when both arguments \code{from} and
\code{to} have their default values \code{NULL}. In this case the
function \code{FUN} will be applied to monthly or quarterly periods.
}
\item{units}{
an optional character string, which allows to overwrite the current
column names of a \code{timeSeries} object. By default \code{NULL}
which means that the column names are selected automatically.
}
\item{format}{
the format specification of the input character vector in POSIX
notation.
}
\item{zone}{
the time zone or financial center where the data were recorded.
}
\item{FinCenter}{
a character value with the the location of the financial center
named as "continent/city", or "city".
}
\item{recordIDs}{
a data frame which can be used for record identification
information. Note, this is not yet handled by the apply functions,
an empty data.frame will be returned.
}
\item{title}{
an optional title string, if not specified the input's data name is
deparsed.
}
\item{documentation}{
optional documentation string, or a vector of character strings.
}
\item{period}{
a character string specifying the rollling period composed by the
length of the period and its unit, e.g. \code{"7d"} represents one
week.
}
\item{\dots}{
arguments passed to other methods.
}
}
\details{
The \code{"timeSeries"} method for \code{apply} extracts the core data
(a matrix) from \code{X} and calls \code{apply}, passing on all the
remaining arguments. If the result is suitable, it converts it to
\code{"timeSeries"}, otherwise returns it as is. \sQuote{Suitable}
here means that it is a matrix or a vector (which is converted to a
matrix) and the number of observations is the same as \code{X}.
Like \code{apply} applies a function to the margins of an array, the
function \code{fapply} applies a function to the time stamps or signal
counts of a financial (therefore the \dQuote{f} in front of the
function name) time series of class \code{"timeSeries"}.
\code{applySeries} takes a \code{"timeSeries"} object as input and
applies \code{FUN} to windows of \code{x}. The windows are specified
by \code{from} and \code{to}, which need to have the same length. Then
\code{from[i], to[i]} specifies the \code{i}-th window. If
\code{time(x)} is a \code{"timeDate"} object, then \code{from} and
\code{to} are converted to \code{"timeDate"} (if they are not already
such objects), otherwise they are converted to integers.
An alternative way to specify the window(s) on which
\code{applySeries} operates is with argument \code{by}. It is used
only if \code{from} and \code{to} are missing or \code{NULL}. \code{by
= "monthly"} or \code{by = "quarterly"} applies \code{FUN} to the data
for each year-month or year-quarter, respectively. By year-month we
mean that there are separate windows for the months in different
years.
The resulting time stamps are the time stamps of the \code{to}
vector. The periods can be regular or irregular, and they can even
overlap.
If \code{from = start(x)} and \code{to = end(x)}, then the function
behaves like \code{apply} on the column margin.
\code{fapply} is the same as \code{applySeries} (in fact, the former
calls the latter), except that the defaults for \code{from} and
\code{to} are \code{start(x)} and \code{end(x)}, respectively. (GNB:
in addition, \code{fapply} throws error if \code{x} is a
\sQuote{signal series}.)
\code{rollDailySeries} rolls a daily 'timeSeries' on a given period.
}
\value{
for \code{rollDailySeries}, an object of class \code{"timeSeries"} with
rolling values, computed from the function \code{FUN}.
}
\examples{
## Percentual Returns of Swiss Bond Index and Performance Index -
LPP <- 100 * LPP2005REC[, c("SBI", "SPI")]
head(LPP, 20)
## Aggregate Quarterly Returns -
applySeries(LPP, by = "quarterly", FUN = colSums)
## Aggregate Quarterly every last Friday in Quarter -
oneDay <- 24*3600
from <- unique(timeFirstDayInQuarter(time(LPP))) - oneDay
from <- timeLastNdayInMonth(from, nday = 5)
to <- unique(timeLastDayInQuarter(time(LPP)))
to <- timeLastNdayInMonth(to, nday = 5)
data.frame(from = as.character(from), to = as.character(to))
applySeries(LPP, from, to, FUN = colSums)
## Alternative Use -
fapply(LPP, from, to, FUN = colSums)
## Count Trading Days per Month -
colCounts <- function(x) rep(NROW(x), times = NCOL(x))
applySeries(LPP, FUN = colCounts, by = "monthly")
## TODO: examples for rollDailySeries()
}
\keyword{chron}
\keyword{ts}
|
