R中使用hill估计的尾部指数

作为我的数据分析（重尾数据）的一部分，我希望计算大约100个公司收益时间序列的尾部（左右）指数。我的数据存储在一个大的zoo对象中，以公司名称作为标题

以下是我目前计算希尔估值器的方法：

returns <- read.zoo("returns.csv", header=TRUE, sep=",", format="%d-%m-%y")
returns_hplots <- lapply(returns, hillPlot)
returns_hill <- sapply(returns_hplots, function(x) x$y)
returns_hill <- sapply(returns_hill, '[', seq(max(sapply(returns_hill,length)))

---

返回值我终于从其他研究人员那里找到了一些示例代码，他们计算了上尾和下尾的hill估计值

以下是我计算两个尾部的hill估值器的代码：

returns <- read.zoo("returns.csv", header=TRUE, sep=",", format="%d-%m-%y")
returns_upper <- lapply(returns, hillPlot, doplot=FALSE) #set doplot to false to speed up calculation
returns_upper <- sapply(returns_upper, function(x) x$y) #extract the hill estimators
returns_upper <- sapply(returns_upper, '[', seq(max(sapply(returns_upper,length))) #create a data frame where each column has equal length (easier analysis)

returns_lower <- lapply(-returns, hillPlot, doplot=FALSE)
returns_lower <- sapply(returns_lower, function(x) x$y)
returns_lower <- sapply(returns_lower, '[', seq(max(sapply(returns_lower,length)))

返回我正在尝试你的代码。语句：返回上
returns_left <- ...
returns_right <- ...

returns <- read.zoo("returns.csv", header=TRUE, sep=",", format="%d-%m-%y")
returns_upper <- lapply(returns, hillPlot, doplot=FALSE) #set doplot to false to speed up calculation
returns_upper <- sapply(returns_upper, function(x) x$y) #extract the hill estimators
returns_upper <- sapply(returns_upper, '[', seq(max(sapply(returns_upper,length))) #create a data frame where each column has equal length (easier analysis)

returns_lower <- lapply(-returns, hillPlot, doplot=FALSE)
returns_lower <- sapply(returns_lower, function(x) x$y)
returns_lower <- sapply(returns_lower, '[', seq(max(sapply(returns_lower,length)))