search

robjhyndman / forecast

Forecasting Functions for Time Series and Linear Models
http://pkg.robjhyndman.com/forecast
1.12k stars 342 forks source link

auto.arima with trace=TRUE values #135

Closed pfv07 closed 9 years ago

pfv07 commented 9 years ago

If the trace=TRUE, this print the model and a value, what is this value ?, because its diferent for the ic values in the best model selected. In the example the value for the best model is 924.795 and ics are: AIC=1029.69 AICc=1029.96 BIC=1037.26

ARIMA(2,1,2)(1,1,1)[12] : Inf ARIMA(0,1,0)(0,1,0)[12] : 954.051 ARIMA(1,1,0)(1,1,0)[12] : 924.795 ARIMA(0,1,1)(0,1,1)[12] : Inf ARIMA(1,1,0)(0,1,0)[12] : 946.442 ARIMA(1,1,0)(2,1,0)[12] : 929.0547 ARIMA(1,1,0)(1,1,1)[12] : Inf ARIMA(1,1,0)(2,1,1)[12] : 929.8475 ARIMA(0,1,0)(1,1,0)[12] : 933.6663 ARIMA(2,1,0)(1,1,0)[12] : 927.8044 ARIMA(1,1,1)(1,1,0)[12] : 926.2389 ARIMA(2,1,1)(1,1,0)[12] : 929.594

Best model: ARIMA(1,1,0)(1,1,0)[12]

Series: dat ARIMA(1,1,0)(1,1,0)[12]

Coefficients: ar1 sar1 0.3392 -0.5716 s.e. 0.0976 0.0899

sigma^2 estimated as 3775: log likelihood=-511.85 AIC=1029.69 AICc=1029.96 BIC=1037.26

robjhyndman commented 9 years ago

If approximation=TRUE, it is the approximated AICc value.