从预测包中使用nnetar函数时出错

时间:2017-06-03 18:19:51

标签: r time-series

我有以下时间序列数据:

is.ts(infy_df_subset_ts)
[1] TRUE
> infy_df_subset_ts
Time Series:
Start = 1 
End = 2573 
Frequency = 1 
   [1] 2272.648 2311.352 2285.300 2275.052 2205.600 2190.100 2164.448 2183.000 2223.400
  [10] 2242.948 2222.200 2205.448 2224.148 2202.352 2254.000 2235.148 2234.052 2241.000
  [19] 2244.400 2247.300 2268.948 2265.552 2269.300 2272.948 2357.052 2373.700 2361.248
  [28] 2351.248 2304.200 2284.648 2382.948 2376.100 2359.948 2310.900 2287.500 2237.700

.................

当我执行nnetar函数时,我收到以下错误:

nnetfit<-nnetar(infy_df_subset_ts,lambda = 0, maxit=150)
Error in nnet.default(x, y, linout = linout, trace = trace, ...) : 
  object 'VR_set_net' not found

我已尝试重新安装预测包,但错误仍然存​​在! 我在这里失踪了什么?

for reproducibilty:ex。取自https://www.otexts.org/fpp/9/3

library(fpp)
fit <- nnetar(sunspotarea)
Error in nnet.default(x, y, linout = linout, trace = trace, ...) : 
  object 'VR_set_net' not found

1 个答案:

答案 0 :(得分:4)

我曾经遇到过这个问题。我要做的是在R里面重新安装包nnet。用以下代码行尝试。

install.packages("nnet")