package - Interpretation auto.arima results in R -


as beginner, trying understand auto.arima function in r forecasting package.

particularly, interested in selection based on information criteria. instance, set ic=c("aicc","aic", "bic"). obtain best fitting model aic, aicc, , bic.

i obtain output value every tested model, e.g. -18661.23 (1,1,1); -18451.12 (1,1,2) etc. if e.g. (1,1,1) selected model lowest output value, value not equal given aic, aicc, or bic.

in simple words, how interpret output value of every model? parallely weighted aic, aicc, , bic?

p.s.: tried understand documentation hard me read.

thank in advance!

as far can tell, "output value" mean value printed when use auto.arima trace=true.

these values aic (or aicc or bic) each model tried. approximation used during search speed things up, value printed may different value returned, calculated without approximation.

the argument ic determines information criterion used. example, setting ic="bic" means bic used in selecting model. default, ic="aicc".

in function definition, argument default value equal vector of values shorthand showing possible values argument can take, first value in vector equal default. in case, function definition contains ic = c("aicc", "aic", "bic") meaning ic can take 1 of 3 values, , default aicc if ic not explicitly passed.


Comments

Popular posts from this blog

Why can rails not find a route created by a helper? -

javascript - jquery or ashx not working -

opencv - DataType<cv::detail::deriv_type>::depth what is it used for -