循环计算价格 returns

Question

我有下面的代码来计算和制作一个data.frame，其中包含我的收盘价 table 的日志返回值。出于某种原因，我只得到 0 个值。不确定问题出在 n 值还是 i 值上。

 number <- c(1:505)
    n <- nrow(data_all)
    for (i in number)
    {returns <- log(data_all[2:n, i]) - log(data_all[1:(n-1), i])
    if (i == number[1]) returns_all <- returns else
      returns_all <- merge(returns_all, returns)
rm(returns)    
}

感谢帮助。

Answer 1

我对你的问题提出 demo.according。代码是

data_all <- as.data.frame(matrix(data = runif(10000,1,50),ncol = 505))
number <- c(1:505)
n <- nrow(data_all)
for (i in number)
{
  returns <- log(data_all[2:n, i]) - log(data_all[1:(n-1), i])
  if (i == number[1]) 
    returns_all <- returns 
  else
    returns_all <- merge(returns_all, returns)
}

我建议你可以按照下面的方式做，

data_all <- as.data.frame(matrix(data = runif(10000,1,50),ncol = 505))
number <- c(1:505)
n <- nrow(data_all)
returns_all <- NULL #I add
for (i in number)
{
  returns <- log(data_all[2:n, i]) - log(data_all[1:(n-1), i])
  returns_all <- cbind(returns_all,returns) #I add
}

希望能帮到你。

Answer 2

因为 log(a) - log(b) = log(a/b) 你可以使用 diff 和 log 函数来计算 returns 的价格系列如下：

对于不止一个系列，我建议将系列转换为xts系列

library("xts")

#if closePx is price series with first column as date with format = 2002-02-12
#specify the corresponding date format , %Y-%m-%d, 
#2002/02/12 => %Y/%m/%d
#closePx_xts = xts(closePx[,-1],order.by = as.Date(closePx[,1],format="%Y-%m-%d")

library("quantmod")

#create environment to store price data
stockNames = c("IBM","MSFT")
dataEnv = new.env()
getSymbols(stockNames, env = dataEnv, from = "2002-01-01", to = "2010-01-01")

#check prices

ls(dataEnv)

head(dataEnv$IBM)
head(dataEnv$MSFT)



#combine closing prices, using Cl function on each list member get closing price 
#for each stock and use merge.xts to combine prices
#if you have missing prices checkout `zoo::na.locf` function
#i.e. na.locf function from zoo package

closePx = do.call(merge.xts,lapply(dataEnv,Cl))

#closePx = Cl(IBM)

#log returns
logRet = diff(log(closePx))


head(logRet)
#              IBM.Close   MSFT.Close
#2002-01-02           NA           NA
#2002-01-03  0.017621634  0.032144867
#2002-01-04  0.015566342 -0.004778131
#2002-01-07 -0.012417504 -0.004946962
#2002-01-08  0.005226094  0.011889353
#2002-01-09 -0.001685453 -0.009703864

使用滞后函数，您可以计算 discrete/arithmetic returns 如下

#discrete returns
discRet = closePx/lag(closePx) - 1


head(discRet)
#              IBM.Close   MSFT.Close
#2002-01-02           NA           NA
#2002-01-03  0.017777811  0.032667094
#2002-01-04  0.015688128 -0.004766734
#2002-01-07 -0.012340725 -0.004934746
#2002-01-08  0.005239774  0.011960312
#2002-01-09 -0.001684034 -0.009656933

还有 PerformanceAnalytics 包以获得更多实用功能

循环计算价格 returns

Loop to calculate price returns

r

return-value