Answers ( 2 )

  1. swaplaw007
    1
    Reply

    when you find the parameters for the ARIMA model from the ACF and PACF plots,
    you get to know how man lags are significant. so you need to test various models
    upto those lags to find out the best model. Suppose, i find that ACF and PACF plots
    are showing significant coefficients upto past 3 lags then i will test various ARIMA models like
    ARIMA(1,0,0) ARIMA(1,0,1) ARIMA(1,0,2) ARIMA(1,0,3)
    ARIMA(2,0,0) ARIMA(2,0,1) ARIMA(2,0,2) ARIMA(2,0,3)
    ARIMA(3,0,0) ARIMA(3,0,1) ARIMA(3,0,2) ARIMA(3,0,3)

    BIC refers to Bayesian Information criteria and is a criterion for a selection
    among the finite set of models. The model with the lowest BIC is the best model.

  2. Shubham Bhatt
    0
    Reply

    Since there is some guesswork in deciding which ARIMA(p,d,q) process to use (i.e. which value to choose for p, q and d), we may need to look at various processes. As for other regression processes, Bayesian Information Criteria (BIC), can be used for this purpose. Generally, the process with the lower BIC value should be selected.
    BIC is an estimate of a function of the posterior probability of a model being true, under a certain Bayesian setup, so that a lower BIC means that a model is considered to be more likely to be the true model.

    Example : Which produces a better model , the ARIMA(2,0,1) model with constant or the ARIMA(2,1,1) model with zero constant.
    Bayesian Information Criteria: BIC = 29.856 (for arima(2,0,1) vs. BIC = 37.267 for arima(2,1,1)
    Thus, we choose the ARIMA(2,0,1) as the better model because the BIC is low for this model.

Leave an answer

Browse
Browse