Question

此代码使我可以使用Dielbold-Li模型构建收益曲线。但是我在代码级别遇到了“构建收益曲线矢量”的问题。应该注意的是，数据的长度与速率相同。您可以生成它。

我正在尝试使用Dielbold-Li模型avec R建立收益率曲线

  ben <- read.csv( "C:\\Users\\BALLA MBAYE DIOP\\Desktop\\Thèse   bureau\\Sortie R\\data\\nex\\ben.csv",sep = ",")

    #########Function to derive Diebold Li Factors################
          di.li.factor.function <- function ( ben, start.date = 0,       set.frequency = 1, tau = c(3,6,12,2*12,3*12,4*12,5*12,6*12,7*12,10*12), lambda = 0.0609)
  # derives the diebold li factors from yield cross section in dataframe yield.data
  # and returns factor, tau is a vector containing the maturity segments of the yield curve in
  #month; start is the starting date of the time series in year +  month/12; frequency = 12 for
  #monthly data; tau is the vector of the maturity structure of the yield data passed in yield.data;
  #lambda is fixed at 0.0609
{
  dimvar <- dim(ben) # writes dimension of yield matrix in variable dimvar(number of
  # rows, number of colums)
  di.li.factor <- mat.or.vec(dimvar[1],3) # initializes matrix (number of rows, number of
  # colums)
  di.li.loading.1 <- mat.or.vec(length(tau),1)
  di.li.loading.2 <- mat.or.vec(length(tau),1)
  di.li.loading.3 <- mat.or.vec(length(tau),1)
  for (loading.count in 1:length(tau))
    # establishes the factor loadings in maturity tau
  {
    di.li.loading.1[loading.count] <- 1
    di.li.loading.2[loading.count] <-
      (1-exp( -lambda*tau[loading.count]))/(lambda*tau[loading.count])
    di.li.loading.3[loading.count] <-
      (1-exp(-lambda *tau[loading.count]))/( lambda *tau[loading.count])-exp(-
                                                                               lambda *tau[loading.count])
  }
  yield.curve <- mat.or.vec(length(tau),1)
  time.count <- 0
  maturity.count <- 0
  for (time.count in 1:dimvar[1])
  {
    for (maturity.count in 1:length(tau))
      #building yield curve vector
    {
      yield.curve[maturity.count] <- ben[time.count, maturity.count]
    }
    #The lm function derives the regression model explainingo the
    #yield.curve vector by the di.li.loading vectors without a constant
    di.li.factor.lm <- lm ( yield.curve ~ -1 +di.li.loading.1
                            +di.li.loading.2
                            +di.li.loading.3)
    #regression coefficients of the regression are written in to the
    #di.li.factor matrix
    di.li.factor[time.count,] <- di.li.factor.lm$coef
  }
  #colums of di.li.factors are labeled according to their interpretation
  colnames(di.li.factor) <- c( "level" , "slope" , "curvature")
  #time series is build by function ts
  di.li.factor <- ts(di.li.factor, start = start.date, end = 10,
                     frequency = set.frequency)
  # di.li.factor multiple time series is returned 
  return (di.li.factor)
}

显示此错误：yield.curve [maturity.count] <-ben [time.count，maturity.count]中的错误：替换的长度为零 ben [time.count，maturity.count]错误：下标超出范围

屈服曲线Diebold-Li

0 个答案: