我正在尝试使用聚合函数将100 Hz的数据转换为1分钟的平均值。但是，当我使用此功能时，1分钟的平均值不正确。数据示例如下。我正在使用以下代码来计算1分钟值。代码没有中断，但是计算不正确。
aggregate(list(X = df$`Gyroscope X`,
Y = df$`Gyroscope Y`,
Z = df$`Gyroscope Z`),
list(minofday = cut(df$Timestamp, "1 min")),mean)
Timestamp Gyroscope X Gyroscope Y Gyroscope Z
2018-07-10T10:25:00.0000000 41.381838 -21.667482 -118.896492
2018-07-10T10:25:00.0100000 48.046268 -12.399903 -110.917976
2018-07-10T10:25:00.0200000 49.102786 -7.36084 -106.485602
2018-07-10T10:25:00.0300000 44.338382 -9.215699 -102.296759
2018-07-10T10:25:00.0400000 34.724123 -11.308594 -96.108404
2018-07-10T10:25:00.0500000 19.622804 -15.225221 -88.122564
2018-07-10T10:25:00.0600000 13.240968 -26.539308 -85.274663
2018-07-10T10:25:00.0700000 13.397218 -31.933596 -80.127568
2018-07-10T10:25:00.0800000 16.333009 -29.663088 -73.027348
2018-07-10T10:25:00.0900000 17.384645 -29.745485 -67.694096
2018-07-10T10:25:00.1000000 16.546632 -30.08423 -67.565922
答案 0 :(得分：1)
假设OP的数据随分钟变化（请注意修改后的数据），这是如何使用基数R和dplyr进行操作：
df$Timestamp <- as.POSIXct(df$Timestamp, format = "%Y-%m-%dT%H:%M:%S")
aggregate(list(X = df$Gyroscope_X,
Y = df$Gyroscope_Y,
Z = df$Gyroscope_Z),
list(minofday = cut(df$Timestamp, "1 min")), mean)
或更简洁的方式：
aggregate(. ~ minofday, mean, data = cbind(setNames(df[,-1], c("X", "Y", "Z")),
minofday = cut(df$Timestamp, "1 min")))
结果：
minofday X Y Z
1 2018-07-10 10:24:00 48.57453 -9.880371 -108.70179
2 2018-07-10 10:25:00 27.78422 -19.314983 -95.13774
3 2018-07-10 10:26:00 16.85883 -29.704286 -70.36072
4 2018-07-10 10:27:00 16.54663 -30.084230 -67.56592
使用lubridate中的summarize_all和dplyr：
library(dplyr)
library(lubridate)
df %>%
mutate(Timestamp = ymd_hms(Timestamp)) %>%
group_by(minofday = cut(Timestamp, "1 min")) %>%
summarize_all(mean) %>%
select(-Timestamp)
结果：
# A tibble: 4 x 4
minofday Gyroscope_X Gyroscope_Y Gyroscope_Z
<fct> <dbl> <dbl> <dbl>
1 2018-07-10 10:24:00 48.6 -9.88 -109.
2 2018-07-10 10:25:00 27.8 -19.3 -95.1
3 2018-07-10 10:26:00 16.9 -29.7 -70.4
4 2018-07-10 10:27:00 16.5 -30.1 -67.6
数据：
df <- read.table(text = " Timestamp Gyroscope_X Gyroscope_Y Gyroscope_Z
2018-07-10T10:25:00.0000000 41.381838 -21.667482 -118.896492
2018-07-10T10:24:00.0100000 48.046268 -12.399903 -110.917976
2018-07-10T10:24:00.0200000 49.102786 -7.36084 -106.485602
2018-07-10T10:25:00.0300000 44.338382 -9.215699 -102.296759
2018-07-10T10:25:00.0400000 34.724123 -11.308594 -96.108404
2018-07-10T10:25:00.0500000 19.622804 -15.225221 -88.122564
2018-07-10T10:25:00.0600000 13.240968 -26.539308 -85.274663
2018-07-10T10:25:00.0700000 13.397218 -31.933596 -80.127568
2018-07-10T10:26:00.0800000 16.333009 -29.663088 -73.027348
2018-07-10T10:26:00.0900000 17.384645 -29.745485 -67.694096
2018-07-10T10:27:00.1000000 16.546632 -30.08423 -67.565922", header = TRUE)
答案 1 :(得分：0)
以下是使用tidyverse的lubridate和dplyr软件包的解决方案：
library(dplyr)
library(lubridate)
df %>%
mutate(day = day(Timestamp),
hour = hour(Timestamp),
min = minute(Timestamp)) %>%
group_by(day, hour, min) %>%
summarise(
`Gyroscope X` = mean(`Gyroscope X`),
`Gyroscope Y` = mean(`Gyroscope Y`),
`Gyroscope Z` = mean(`Gyroscope Z`)
)
答案 2 :(得分：0)
由于您正在处理时间戳，因此xts软件包具有很多可以帮助您的功能。要汇总时间戳记，period.apply可以为您提供帮助。 endpoints部分可以将数据从微秒累积到数年。
# don't load the timestamp column that one goes to the order.by part
df1_xts <- xts(df1[, -1], order.by = df1$Timestamp)
# roll up to seconds.
period.apply(df1_xts, endpoints(df1_xts, on = "mins"), colMeans)
Gyroscope_X Gyroscope_Y Gyroscope_Z
2018-07-10 10:25:00 28.55624 -20.46759 -90.59249
如果您的时间戳列还不是日期时间对象，则可以使用以下方法：
df1$Timestamp <- strptime(df1$Timestamp, format = "%Y-%m-%dT%H:%M:%OS")
数据：
df1 <- structure(list(Timestamp = structure(list(sec = c(0, 0.01, 0.02,
0.03, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.1), min = c(25L,
25L, 25L, 25L, 25L, 25L, 25L, 25L, 25L, 25L, 25L), hour = c(10L,
10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L), mday = c(10L,
10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L, 10L), mon = c(6L,
6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L, 6L), year = c(118L, 118L,
118L, 118L, 118L, 118L, 118L, 118L, 118L, 118L, 118L), wday = c(2L,
2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L), yday = c(190L, 190L,
190L, 190L, 190L, 190L, 190L, 190L, 190L, 190L, 190L), isdst = c(1L,
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L), zone = c("CEST", "CEST",
"CEST", "CEST", "CEST", "CEST", "CEST", "CEST", "CEST", "CEST",
"CEST"), gmtoff = c(NA_integer_, NA_integer_, NA_integer_, NA_integer_,
NA_integer_, NA_integer_, NA_integer_, NA_integer_, NA_integer_,
NA_integer_, NA_integer_)), class = c("POSIXlt", "POSIXt")),
Gyroscope_X = c(41.381838, 48.046268, 49.102786, 44.338382,
34.724123, 19.622804, 13.240968, 13.397218, 16.333009, 17.384645,
16.546632), Gyroscope_Y = c(-21.667482, -12.399903, -7.36084,
-9.215699, -11.308594, -15.225221, -26.539308, -31.933596,
-29.663088, -29.745485, -30.08423), Gyroscope_Z = c(-118.896492,
-110.917976, -106.485602, -102.296759, -96.108404, -88.122564,
-85.274663, -80.127568, -73.027348, -67.694096, -67.565922
)), row.names = c(NA, -11L), class = "data.frame")