R：按团队计算平均值

我有以下格式的足球成绩数据（数千次观察）：

---

加上其他变量。“价值”是团队，pts是最终结果（赢/输/平）作为数值。我试图添加一个新的变量，它是该行球队在过去X场比赛中该值的平均值。我如何做到这一点而不使用一些可怕的循环

尝试使用stats中的

ave

函数

Trt <- gl(n=2, k=3, length=2*3, labels =c("A", "B"))
Y <- 1:6
Data <- data.frame(Trt, Y)
 Data
  Trt Y
1   A 1
2   A 2
3   A 3
4   B 4
5   B 5
6   B 6
Data$TrtMean <- ave(Y, Trt, FUN=mean)
Data
  Trt Y TrtMean
1   A 1       2
2   A 2       2
3   A 3       2
4   B 4       5
5   B 5       5
6   B 6       5

Trt看看

使用zoo软件包和
rollmean
以及plyr软件包的
ddply
：

library(zoo)
library(plyr)
dat <- data.frame(value=letters[1:5], pts=sample(c(0, 0.5, 1), 50, replace=T))
ddply(dat, .(value), summarise, rollmean(pts, k=5, align='right'))

使用
tapply
可以非常有效地完成此操作。我用随机的分数和日期复制了球队的比赛，在某种程度上改变了你的数据。这取最近2场比赛的平均值，如
tail
函数中所指定

# create some data
d <- structure(list(Div = structure(rep(1L, 33), .Label = " E0", 
  class = "factor"), date = structure(c(15013, 14990, 14996, 15001, 14995, 15006, 
  15020, 15032, 15023, 15022, 15015, 15016, 15034, 14994, 14986, 14998, 14982, 
  14979, 14980, 15016, 15031, 15013, 15031, 14999, 15025, 14978, 15007, 15026, 
  14992, 14997, 15023, 14986, 15028), class = "Date"), 
  value = structure(c(3L, 4L, 5L, 7L, 8L, 11L, 9L, 10L, 6L, 1L, 2L, 3L, 4L, 5L, 
  7L, 8L, 11L, 9L, 10L, 6L, 1L, 2L, 3L, 4L, 5L, 7L, 8L, 11L, 9L, 10L, 6L, 1L, 
  2L), .Label = c("Arsenal", "Aston Villa", "Blackburn", "Fulham", "Liverpool",
  "Man City", "Newcastle", "QPR", "Stoke", "West Brom", "Wigan"), 
  class = "factor"), pts = c(0.5, 0.5, 0.5, 1, 1, 1, 1, 0, 1, 0.5, 0, 1, 1, 1, 1, 
  0.5, 0.5, 0, 0.5, 0.5, 0, 0, 0, 1, 0, 0, 0.5, 0, 1, 0, 0.5, 0.5, 0.5)), 
  .Names = c("Div", "date", "value", "pts"), row.names = c(NA, 33L), 
  class = "data.frame")

# sort rows by date
d2 <- d[order(d$date),]
# mean of all games
tapply(d2$pts, d2$value, mean)
# mean of last 2 games
tapply(d2$pts, d2$value, function(x) mean(tail(x, 2)))

# To tidy up the output, you could use simplify=FALSE and do.call(rbind, x):
# e.g., mean of last 2 games:
do.call(rbind, tapply(d2$pts, d2$value, function(x) mean(tail(x, 2)), 
  simplify=F))

            [,1]
Arsenal     0.25
Aston Villa 0.25
Blackburn   0.50
Fulham      1.00
Liverpool   0.25
Man City    0.75
Newcastle   1.00
QPR         0.50
Stoke       1.00
West Brom   0.00
Wigan       0.50

#创建一些数据
d请提供一些示例代码，这使OP更加清晰。@PaulHiemstra：我添加了一个示例。是否有一种简单的方法来修改此代码以满足问题的要求？事实上，
aggregate
只需一步，例如
聚合（d2$pts，list（d2$value），函数（x）均值（tail（x，2））
ddply(dat, .(value), summarise, rollmean(pts, k=5, fill=NA, align='right'))

# create some data
d <- structure(list(Div = structure(rep(1L, 33), .Label = " E0", 
  class = "factor"), date = structure(c(15013, 14990, 14996, 15001, 14995, 15006, 
  15020, 15032, 15023, 15022, 15015, 15016, 15034, 14994, 14986, 14998, 14982, 
  14979, 14980, 15016, 15031, 15013, 15031, 14999, 15025, 14978, 15007, 15026, 
  14992, 14997, 15023, 14986, 15028), class = "Date"), 
  value = structure(c(3L, 4L, 5L, 7L, 8L, 11L, 9L, 10L, 6L, 1L, 2L, 3L, 4L, 5L, 
  7L, 8L, 11L, 9L, 10L, 6L, 1L, 2L, 3L, 4L, 5L, 7L, 8L, 11L, 9L, 10L, 6L, 1L, 
  2L), .Label = c("Arsenal", "Aston Villa", "Blackburn", "Fulham", "Liverpool",
  "Man City", "Newcastle", "QPR", "Stoke", "West Brom", "Wigan"), 
  class = "factor"), pts = c(0.5, 0.5, 0.5, 1, 1, 1, 1, 0, 1, 0.5, 0, 1, 1, 1, 1, 
  0.5, 0.5, 0, 0.5, 0.5, 0, 0, 0, 1, 0, 0, 0.5, 0, 1, 0, 0.5, 0.5, 0.5)), 
  .Names = c("Div", "date", "value", "pts"), row.names = c(NA, 33L), 
  class = "data.frame")

# sort rows by date
d2 <- d[order(d$date),]
# mean of all games
tapply(d2$pts, d2$value, mean)
# mean of last 2 games
tapply(d2$pts, d2$value, function(x) mean(tail(x, 2)))

# To tidy up the output, you could use simplify=FALSE and do.call(rbind, x):
# e.g., mean of last 2 games:
do.call(rbind, tapply(d2$pts, d2$value, function(x) mean(tail(x, 2)), 
  simplify=F))

            [,1]
Arsenal     0.25
Aston Villa 0.25
Blackburn   0.50
Fulham      1.00
Liverpool   0.25
Man City    0.75
Newcastle   1.00
QPR         0.50
Stoke       1.00
West Brom   0.00
Wigan       0.50