Question

在R中，如果我有这样的时间序列：

datetime          value
2014-12-01 00:00  100
2014-12-01 01:00  111
2014-12-01 02:00  98
2014-12-01 03:00  127
2014-12-01 04:00  114
2014-12-01 05:00  105
...

当这个系列的累积达到某个值时，如何获得插值日期时间？像这样：

somefunction(data, 250) // returns 2014-12-01 01:23

Answer 1

您可以使用approx：

执行此操作

Data <- structure(list(datetime = structure(c(1417413600, 1417417200, 
  1417420800, 1417424400, 1417428000, 1417431600), class = c("POSIXct", 
  "POSIXt"), tzone = ""), value = c(100L, 111L, 98L, 127L, 114L, 105L)),
  .Names = c("datetime", "value"), row.names = c(NA, -6L),
  class = "data.frame")
itime <- .POSIXct(approx(cumsum(Data$value), Data$datetime, 250)$y,
                  tz=attr(Data$datetime, "tzone"))

Answer 2

首先，确保您的datetime是一个不错的POSIXct对象，如下所示：

foo <- structure(list(datetime = structure(c(1417388400, 1417392000, 
1417395600, 1417399200, 1417402800, 1417406400), class = c("POSIXct", 
"POSIXt"), tzone = ""), value = c(100, 111, 98, 127, 114, 105
)), .Names = c("datetime", "value"), row.names = c(NA, -6L), class = "data.frame")
foo
             datetime value
1 2014-12-01 00:00:00   100
2 2014-12-01 01:00:00   111
3 2014-12-01 02:00:00    98
4 2014-12-01 03:00:00   127
5 2014-12-01 04:00:00   114
6 2014-12-01 05:00:00   105

然后，approxfun是你的朋友：

ff <- with(foo,approxfun(x=cumsum(value),y=datetime))
as.POSIXct(ff(250),origin=as.POSIXct("1970-01-01 01:00:00"))
[1] "2014-12-01 01:23:52 CET"

如何从累积时间序列中获取阈值

2 个答案: