如何将时间序列原始日期保存在rugarch R包中？

Question

在R中，我有一个时间序列对象（ts）

library(rugarch)
library(TSA)

observations <- read.csv("financial_data.csv", header = TRUE)
start = c(1999, as.numeric(format(as.Date("1999-05-19"), "%j")))
end = c(2003, as.numeric(format(as.Date("2003-10-04"), "%j")))
ts.obj <- ts(observations, start=start, end=end, frequency=365) # daily observations


model.spec <- rugarch:::ugarchspec(
  variance.model = list(
    model = "sGARCH", 
    garchOrder = c(1, 1),
    submodel = NULL,
    external.regressors = NULL,
    variance.targeting = FALSE), 
  mean.model = list(
    armaOrder = c(1, 1), 
    include.mean = FALSE,
    external.regressors = NULL),
    distribution.model = "norm"
  )

model.fit <- rugarch:::ugarchfit(spec=model.spec,
                                data=ts.obj, # the ts object 
                                solver.control = list(trace=0))

forc = rugarch:::ugarchforecast(model.fit, n.ahead = 10)
rugarch:::plot(forc)

当我将此时间序列传递给 rugarch ::: ugarchfit 函数时，时间轴从1970-01-01开始，而不是1999-05-19，因此预测时间轴似乎在过去发生了偏移以及错误的年份，月份和日期。

如何将原始日期保留在预测图中？