我一直在研究R中的一些时间序列数据，由于连续点之间存在一些差异，因此我想通过计算移动平均值来平滑数据。
我很难找到一个内置函数来执行此操作，但是遇到了迪迪埃·鲁丁（Didier Ruedin）的博客文章 ，其中描述了执行该任务所需的以下函数：
mav <- function(x,n=5){filter(x,rep(1/n,n), sides=2)}我尝试插入一些数字以了解其工作原理：
> mav(c(4,5,4,6), 3)
Time Series:
Start = 1
End = 4
Frequency = 1
[1] NA 4.333333 5.000000 NA在这里，我试图进行滚动平均，并考虑到最后三个数字，因此我希望只返回两个数字– 4.333333和5 –如果要使用NA值，我认为它们应该是在开始时序列。
实际上，事实证明这是'sides'参数所控制的：
sides
for convolution filters only. If sides = 1 the filter coefficients are for past values only; if sides = 2 they
are centred around lag 0. In this case the length of the filter should be odd, but if it is even, more of the
filter is forward in time than backward.因此，在我们的“ mav”函数中，滚动平均值看起来是当前值的两侧，而不仅仅是过去的值。 我们可以对其进行调整以获得所需的行为：
mav <- function(x,n=5){filter(x,rep(1/n,n), sides=1)}> mav(c(4,5,4,6), 3)
Time Series:
Start = 1
End = 4
Frequency = 1
[1] NA NA 4.333333 5.000000NA值对于我们想要进行的任何绘制都很烦人，因此让我们摆脱它们：
> na.omit(mav(c(4,5,4,6), 3))
Time Series:
Start = 3
End = 4
Frequency = 1
[1] 4.333333 5.000000至此，我注意到Didier在注释中引用了zoo程序包，并且它具有内置功能来处理所有这一切：
> library(zoo)
> rollmean(c(4,5,4,6), 3)
[1] 4.333333 5.000000我还意识到我可以使用'ls'函数列出软件包中的所有函数，因此下次我需要做与时间序列相关的事情时，我将扫描Zoo的函数列表–可能已经有一个函数！
> ls("package:zoo")
[1] "as.Date" "as.Date.numeric" "as.Date.ts"
[4] "as.Date.yearmon" "as.Date.yearqtr" "as.yearmon"
[7] "as.yearmon.default" "as.yearqtr" "as.yearqtr.default"
[10] "as.zoo" "as.zoo.default" "as.zooreg"
[13] "as.zooreg.default" "autoplot.zoo" "cbind.zoo"
[16] "coredata" "coredata.default" "coredata<-"
[19] "facet_free" "format.yearqtr" "fortify.zoo"
[22] "frequency<-" "ifelse.zoo" "index"
[25] "index<-" "index2char" "is.regular"
[28] "is.zoo" "make.par.list" "MATCH"
[31] "MATCH.default" "MATCH.times" "median.zoo"
[34] "merge.zoo" "na.aggregate" "na.aggregate.default"
[37] "na.approx" "na.approx.default" "na.fill"
[40] "na.fill.default" "na.locf" "na.locf.default"
[43] "na.spline" "na.spline.default" "na.StructTS"
[46] "na.trim" "na.trim.default" "na.trim.ts"
[49] "ORDER" "ORDER.default" "panel.lines.its"
[52] "panel.lines.tis" "panel.lines.ts" "panel.lines.zoo"
[55] "panel.plot.custom" "panel.plot.default" "panel.points.its"
[58] "panel.points.tis" "panel.points.ts" "panel.points.zoo"
[61] "panel.polygon.its" "panel.polygon.tis" "panel.polygon.ts"
[64] "panel.polygon.zoo" "panel.rect.its" "panel.rect.tis"
[67] "panel.rect.ts" "panel.rect.zoo" "panel.segments.its"
[70] "panel.segments.tis" "panel.segments.ts" "panel.segments.zoo"
[73] "panel.text.its" "panel.text.tis" "panel.text.ts"
[76] "panel.text.zoo" "plot.zoo" "quantile.zoo"
[79] "rbind.zoo" "read.zoo" "rev.zoo"
[82] "rollapply" "rollapplyr" "rollmax"
[85] "rollmax.default" "rollmaxr" "rollmean"
[88] "rollmean.default" "rollmeanr" "rollmedian"
[91] "rollmedian.default" "rollmedianr" "rollsum"
[94] "rollsum.default" "rollsumr" "scale_x_yearmon"
[97] "scale_x_yearqtr" "scale_y_yearmon" "scale_y_yearqtr"
[100] "Sys.yearmon" "Sys.yearqtr" "time<-"
[103] "write.zoo" "xblocks" "xblocks.default"
[106] "xtfrm.zoo" "yearmon" "yearmon_trans"
[109] "yearqtr" "yearqtr_trans" "zoo"
[112] "zooreg"善于交际，分享！
翻译自: https://www.javacodegeeks.com/2014/09/r-calculating-rolling-or-moving-averages.html