我有以下数据集与每日收集的数据.
R> toydata
date group coef
1 2011-11-04 23:59:59 1 2.32
2 2011-11-05 23:59:59 1 2.34
3 2011-11-06 23:59:59 1 2.46
4 2011-11-07 23:59:59 1 2.68
5 2011-11-04 23:59:59 2 2.17
6 2011-11-05 23:59:59 2 1.90
7 2011-11-06 23:59:59 2 2.13
8 2011-11-07 23:59:59 2 2.52
Everythings似乎工作得很好:
R> toydata$date
[1] "2011-11-04 23:59:59 EST" "2011-11-05 23:59:59 EST"
[3] "2011-11-06 23:59:59 EST" "2011-11-07 23:59:59 EST"
[5] "2011-11-04 23:59:59 EST" "2011-11-05 23:59:59 EST"
[7] "2011-11-06 23:59:59 EST" "2011-11-07 23:59:59 EST"
和
R> format(toydata$date, …我正在分析一个数据集,其中数据聚集在几个组(区域中的城镇)中.数据集如下所示:
R> df <- data.frame(x = rnorm(10),
y = 3*rnorm(x),
groups = factor(sample(c('0','1'), 10, TRUE)))
R> head(df)
x y groups
1 -0.8959 1.54 1
2 -0.1008 -2.73 1
3 0.4406 0.44 0
4 0.0683 1.62 1
5 -0.0037 -0.20 1
6 -0.8966 -2.34 0
我希望我的LM()估计占同类相关的群体,并为此我使用一个函数cl(),它接受一个lm()和(原返回可靠的集群的协方差矩阵这里):
cl <- function(fm, cluster) {
library(sandwich)
M <- length(unique(cluster))
N <- length(cluster)
K <- fm$rank
dfc <- (M/(M-1))*((N-1)/(N-K-1))
uj <- apply(estfun(fm), 2, function(x) tapply(x, cluster, sum));
vcovCL <- dfc …