这些时间序列相关性在R中意味着什么？

Question

假设我的两个时间序列数据集如下：

temp <- matrix(c(12,8,9,11,10,3,7,2,4), byrow = TRUE, ncol = 3)
rownames(temp) <- c("Denver", "LA", "NY")
colnames(temp) <- c("day-1", "day-2", "day-3")
humidity <- matrix(c(70,67,34,45,56,47,68,74,36), byrow = TRUE, ncol = 3)
colnames(humidity)<- c("day-1", "day-2", "day-3")
rownames(humidity)<- c("Denver", "LA", "NY")


> temp
       day-1 day-2 day-3
Denver    12     8     9
LA        11    10     3
NY         7     2     4


> humidity
       day-1 day-2 day-3
Denver    70    67    34
LA        45    56    47
NY        68    74    36



> cor(humidity)
           day-1      day-2      day-3
day-1  1.0000000  0.8924544 -0.9974590
day-2  0.8924544  1.0000000 -0.8580461
day-3 -0.9974590 -0.8580461  1.0000000

> cor(temp)
          day-1     day-2     day-3
day-1 1.0000000 0.9078413 0.5291067
day-2 0.9078413 1.0000000 0.1245339
day-3 0.5291067 0.1245339 1.0000000


> cor(humidity, temp)
           day-1      day-2      day-3
day-1 -0.2584615 -0.6397084  0.6829690
day-2 -0.6664738 -0.9176629  0.2799863
day-3  0.1889822  0.5833240 -0.7332730

我不知道如何解释这个时间序列？

1-哪个cor显示变量是否相关？

2-为什么城市不被忽视？

3-在第三个cor(humidity, temp)中，主对角线值不是cor矩阵中的一个。怎么会这样?

Answer 1

第三矩阵(cor(humidity, temp))的(i，j)第四元素显示了humidity的第一列与temp的jth列之间的相关性。说明：

cor(c(12,11,7),c(70,45,68)) ## day-1 humidities vs day-1 temps (m[1,1])
cor(c(8,10,2), c(70,45,68)) ## day-2 humidities vs day-1 temps (m[2,1])

如果您设置

cc <- cor(humidity, temp)
names(dimnames(cc)) <- c("humidity", "temp")