CODEWITHSUNDEEP
rseasonal-adjustment
Thomas Rosa
Thomas Rosa

Reputation: 742

What is causing X-13 times series seasonalization to fail for this specific series (using the R seas package)?

I am seasonalizing a ton of times series data and all of them work fine except this one. Can someone help me determine why it fails and what would be a good way for me to address this?

library(seasonal)
data = c(10, 8, 7, 16, 11, 20, 14, 14, 17, 13, 10, 9,
         10, 9, 10, 18, 20, 13, 17, 15, 20, 18, 13, 5,
         13, 14, 14, 14, 12, 10, 32, 23, 7, 11, 8, 8,
         15, 14, 12, 11, 13, 18, 18, 13, 9, 15, 9, 10,
         13, 8, 15, 11, 15, 17, 13, 11, 14, 20, 11, 7,
         11, 8, 10, 18, 19, 19, 23, 7, 16, 11, 14, 4,
         14, 16, 14, 14, 24, 15, 17, 21, 14, 24)
series = ts(data, frequency = 12, start = c(2017, 1))
# print(series)
# plot(series)
seas(series)

Error message:

Error: X-13 run failed

Errors:
- The covariance matrix of the ARMA parameters is singular; cannot compute
  t-statistics for the ARMA parameters. Program error(s) halt execution for
  C:\Users\THOMAS~1\AppData\Local\Temp\RtmpQHZhsm\x13391c5b19350e/iofile.spc

Notes:
- IEEE_INVALID_FLAG IEEE_DIVIDE_BY_ZERO IEEE_UNDERFLOW_FLAG IEEE_DENORMAL

If you remove the last value in the series, or make it lower than 20 or higher than 27, it works again.

Upvotes: 0

Views: 436

Answers (1)

Thomas Rosa
Thomas Rosa

Reputation: 742

The suggestions and explanations here can resolve a situation like this. Which solution is best will depend on what kind of data this is. The options listed are:

seas(series, outlier = NULL)
seas(series, arima.model = c(0, 1, 1, 0, 1, 1))
seas(series, regression.aictest = NULL)

Upvotes: 1

Related Questions