ssa()时间序列填充R

Question

我有一个关于时间序列的缺口的问题，使用R包{Rssa}。该示例系列的长度为334，并且包含多个NAs。

b1 <- as.vector(B1.med[1,])
> str(b1)
 num [1:334] 0.0171 0.0331 0.014 0.0371 0.0149 0.0179 NA NA 0.0293 0.0096 ...

我想使用igapfill {Rssa}来填充我的NAs。因为igapfill想要一个SSA形状的对象，所以我必须首先做一个ssa，其中出现了一个问题：

> my.L = length(b1)%/%2
> s1.gap <- ssa(b1,L=my.L)
Error in ssa(b1, L = my.L) : 
  Nothing to decompose: the given field shape is empty
In addition: Warning messages:
1: In eval(expr, envir, enclos) :
  Some field elements were not covered by shaped window. 292 elements will be ommited
2: In eval(expr, envir, enclos) :
  Nothing to decompose: the given field shape is empty

帮助建议(默认情况下)窗口长度L大约是数据长度的一半，这对我不起作用。只有在将L设置为23之后，我才不会收到此错误消息，但仍然会出现警告：

> s2.gap <- ssa(b1,L=23)
Warning message:
In eval(expr, envir, enclos) :
  Some field elements were not covered by shaped window. 269 elements will be ommited

> summary(s2.gap) #for L=23

Call:
ssa(x = b1, L = 23)

Series length: 334, Window length: 23,  SVD method: eigen
Special triples:  0

Computed:
Eigenvalues: 1, Eigenvectors: 1,    Factor vectors: 0

Precached: 0 elementary series (0 MiB)

Overall memory consumption (estimate): 0.008934 MiB

但我不会得到超过1个特征向量，直到我进一步减少(仍然是错误消息)。

> summary(s2.gap)  #for L=15

Call:
ssa(x = b1, L = 15)

Series length: 334, Window length: 15,  SVD method: eigen
Special triples:  0

Computed:
Eigenvalues: 15,    Eigenvectors: 15,   Factor vectors: 0

Precached: 0 elementary series (0 MiB)

Overall memory consumption (estimate): 0.01062 MiB

我不太明白我的问题是什么，谷歌等也帮不了我太多忙。我非常感谢您的建议:-)