R 时间序列 10 分钟间隔

Question

我有一个大型数据集（临时），几年来每十分钟进行一次测量。 “值”列也包含 NA。 如何创建正确的时间序列对象？我尝试使用 ts () 函数，但我只是没能找到适合我的数据集的频率。 之后我想进行趋势/周期性/季节性的相关性和分析。

> dput(head(temp))
structure(list(TS = structure(c(874659600, 874660200, 874660800, 
874661400, 874662000, 874662600), class = c("POSIXct", "POSIXt"
), tzone = "UTC"), value = c(2.77268641140869, 2.77268641140869, 
2.77268641140869, 2.77268641140869, 2.71474135464008, 2.71474135464008
)), row.names = c(NA, 6L), class = "data.frame")

Answer 1

如果需要改变频率，需要您自行决定或根据现有情况决定。您可以使用

na.approx()

包中的

zoo

来插入“值”并消除 NA。

#if we change the frequency to 1 minute
temp%<>%merge(data.frame('TS' =seq(min(temp$TS), max(temp$TS),  by = '1 min')), all=T)
temp$value = zoo::na.approx(temp$value)
                    TS    value
1  1997-09-19 09:00:00 2.772686
2  1997-09-19 09:01:00 2.772686
3  1997-09-19 09:02:00 2.772686
4  1997-09-19 09:03:00 2.772686
5  1997-09-19 09:04:00 2.772686
6  1997-09-19 09:05:00 2.772686
7  1997-09-19 09:06:00 2.772686
8  1997-09-19 09:07:00 2.772686
9  1997-09-19 09:08:00 2.772686
10 1997-09-19 09:09:00 2.772686
...
31 1997-09-19 09:30:00 2.772686
32 1997-09-19 09:31:00 2.766892
33 1997-09-19 09:32:00 2.761097
34 1997-09-19 09:33:00 2.755303
35 1997-09-19 09:34:00 2.749508
36 1997-09-19 09:35:00 2.743714
37 1997-09-19 09:36:00 2.737919
38 1997-09-19 09:37:00 2.732125
39 1997-09-19 09:38:00 2.726330
40 1997-09-19 09:39:00 2.720536
41 1997-09-19 09:40:00 2.714741
42 1997-09-19 09:41:00 2.714741
43 1997-09-19 09:42:00 2.714741
44 1997-09-19 09:43:00 2.714741
45 1997-09-19 09:44:00 2.714741
46 1997-09-19 09:45:00 2.714741
47 1997-09-19 09:46:00 2.714741
48 1997-09-19 09:47:00 2.714741
49 1997-09-19 09:48:00 2.714741
50 1997-09-19 09:49:00 2.714741
51 1997-09-19 09:50:00 2.714741