Error : No suitable ARIMA model found using tso

Question

I am working on a model to forecast demand for a product and wanted to remove outliers from my time series. I came across the tso function and decided to give it a try. The initial results were encouraging and hence decided to pursue it. However, when I run it across all the different time series I often come across an error. Providing a description of the code and error below.

Code:

data<-c(2780,1580,2800,3320,4720,2860,3440,2940,3720,3360,3380,3160,2220,2560,3260, 3140,1920,3340,3060,1960,2800,1520,1920,1800,2100,2240,2300,1560,820,2160)

dataTS<-ts(data, frequency = 12)

dataTS.out<-tso(dataTS, tsmethod = "auto.arima", args.tsmethod = list(ic = "bic"), maxit = 5)

Error Message

Error in auto.arima(x = c(2780, 1580, 2800, 3320, 4720, 2860, 3440, 2940, : No suitable ARIMA model found

What am I doing wrong?

Thanking in advance

Answer 1

It seems to be omitting an argument in the list. This worked for me.

library(tsoutliers)
dataTS.out <- tso(dataTS, tsmethod = "auto.arima", args.tsmethod = list(allowdrift = FALSE, ic = "bic"), maxit = 5)
dataTS.out