Question

我正在尝试从具有2个列的简单数据矩阵计算对数返回。我的以下代码给了我全零，我在这里错过了什么？

dataMatrix<-as.matrix(data[,2:ncol(data)])  #taking data from dataframe 
Log_returns<-diff(log(dataMatrix), lag=1)

View(log_returns)

这是数据的前几行

        Dates   APPL  AMZN
1   7-01-2005 4.9464 42.32
2  2005-01-14 5.0143 44.55
3  2005-01-21 5.0350 41.16
4  2005-01-28 5.2843 42.22
5  2005-02-04 5.6314 35.72
6  2005-02-11 5.8007 35.78
7  2005-02-18 6.2007 35.31
8  2005-02-25 6.3564 34.99
9  2005-03-04 6.1157 35.85
10 2005-03-11 5.7529 34.75

Answer 1

我不确定您要计算的内容。您是要计算每一行APPL和AMZN的对数转换后的值之差吗？

Log_returns <- apply(log(data[, 2:ncol(data)]), 1, diff)
#       1        2        3        4        5        6        7        8
#2.146600 2.184318 2.101053 2.078154 1.847353 1.819410 1.739504 1.705600
#       9       10
#1.768484 1.798475

或者保存在新列中

data$Log_return <- apply(log(data[, 2:ncol(data)]), 1, diff)
data
#        Dates   APPL  AMZN Log_return
#1   7-01-2005 4.9464 42.32   2.146600
#2  2005-01-14 5.0143 44.55   2.184318
#3  2005-01-21 5.0350 41.16   2.101053
#4  2005-01-28 5.2843 42.22   2.078154
#5  2005-02-04 5.6314 35.72   1.847353
#6  2005-02-11 5.8007 35.78   1.819410
#7  2005-02-18 6.2007 35.31   1.739504
#8  2005-02-25 6.3564 34.99   1.705600
#9  2005-03-04 6.1157 35.85   1.768484
#10 2005-03-11 5.7529 34.75   1.798475

也许您的意思是计算每个APPL和AMZN的连续日期之间对数转换后的值的差异

apply(log(data[, 2:ncol(data)]), 2, diff)
#           APPL         AMZN
#2   0.013633791  0.051352366
#3   0.004119696 -0.079145243
#4   0.048326634  0.025427131
#5   0.063617919 -0.167183286
#6   0.029620521  0.001678322
#7   0.066683589 -0.013222868
#8   0.024799990 -0.009103904
#9  -0.038602944  0.024281261
#10 -0.061155160 -0.031163995

样本数据

data <- read.table(text =
    "        Dates   APPL  AMZN
1   7-01-2005 4.9464 42.32
2  2005-01-14 5.0143 44.55
3  2005-01-21 5.0350 41.16
4  2005-01-28 5.2843 42.22
5  2005-02-04 5.6314 35.72
6  2005-02-11 5.8007 35.78
7  2005-02-18 6.2007 35.31
8  2005-02-25 6.3564 34.99
9  2005-03-04 6.1157 35.85
10 2005-03-11 5.7529 34.75", header = T)

使用Diff函数计算日志返回

1 个答案:

样本数据