is there a possibility to check the coefficients of an ARIMA Model with the ARIMA Learner Node? On the one hand a model is great if it has low information criteria and high log-likelihood and on the other hand if all coefficients are significant. It would be great if there is a way to check this.
The coefficients, log likelihood, and AIC/BIC statistics are shown in the middle output of the learner node. The p-values of the coefficients are not given there, but the standard errors are.
Regarding the significance of the coefficients: If the coefficient estimate is more than two (1,96) standard errors away from 0 (significance 95%) then the coefficient is significant? For example: The coefficient is 1 and the standard error is 0,2; then the coefficient is significant because 1 is bigger than 0,4 (0,2*2).
In general, dividing the coefficient value by the standard deviation gives the test statistics z. z follows the standard normal distribution and gives you the p-value, 0.05 for z=1.96 for example.