CODEWITHSUNDEEP
rforecastingarima
RParker
RParker

Reputation: 1

R forecast::auto.arima() not including seasonal differencing

While trying to create ARIMA models for half-hour time series data, auto.arima would not automatically include seasonality, despite apparent (daily) seasonality. When specifying D = 1 when calling auto.arima (as instructed here), the ARIMA model returns the following error:

Error in if (length(dx) > ncol(dxreg)) lm(dx ~ dxreg - 1, na.action = na.omit) else list(rank = 0L) : 
  argument is of length zero

Have included a minimal reproducible example below

structure(c(860.110666666667, 0, 0, 0, 0, 0, 0, 0, 1795.942, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
1233.68133333333, 0, 0, 0, 0, 2460.29066666667, 1910.628, 3026.91, 
1313.402, 587.111666666667, 1389.328, 0, 425.473666666667, 2676.30433333333, 
359.064, 598.866666666667, 0, 0, 0, 0, 0, 0, 826.639333333333, 
0, 0, 0, 0, 0, 525.482333333333, 600.823333333333, 0, 1293.36533333333, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 383.521333333333, 
0, 0, 0, 0, 0, 0, 739.561, 3241.267, 2820.05166666667, 3262.343, 
1779.022, 1406.778, 631.806666666667, 3261.663, 25.2043333333333, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1120.446, 931.430666666667, 329.675, 
167.672, 2957.59733333333, 0, 0, 0, 0, 0, 0, 486.451666666667, 
1210.21233333333, 0, 1192.211, 0, 0, 0, 0, 0, 0, 1899.427, 0, 
0, 0, 0, 0, 0, 717.466333333333, 0, 0, 0, 2421.93233333333, 3231.43766666667, 
1160.77066666667, 2315.53966666667, 3252.93833333333, 3210.40633333333, 
1805.097, 1445.225), .Tsp = c(1505352600, 1505352600.00816, 17520
), class = "ts")

aa <- forecast::auto.arima(dfts, D = 1)

Upvotes: 0

Views: 403

Answers (1)

Rob Hyndman
Rob Hyndman

Reputation: 31800

You define a time series of length 144 with frequency equal to 17520 and then ask for a seasonal difference -- i.e., a difference between observations that are 17520 time periods apart -- which can't be done.

The error message is not very helpful -- I'll get that fixed.

Upvotes: 1

Related Questions