1

Is it possible to calculate an R squared value from an ARIMA model in R?

This is the output given from summary(model)

edit: I am worried about the biases associated with MAPE and other percentage errors. The quantities I'm predicting are relatively small so I feel that finding R2, correlation or some sort of other metric might be a better indicator.

sarahbarnes
  • 103
  • 2
  • 7
  • 1
    Try `summary(model)$r.squared`. If that doesn't work, take a look at `str(summary(model))` and you can find what you're looking for. – mickey Dec 04 '18 at 19:24
  • Regarding your screenshot: an ARIMA(0,0,0) is just a mean, so this model is just estimating the mean. – Frank Dec 05 '18 at 19:14

1 Answers1

1

Once you have ARMA errors, it is not a simple linear regression any more.

Zeeshan
  • 1,208
  • 1
  • 14
  • 26
  • Yes, but I am wondering if there is any way to calculate either R2 or correlation. I am currently using MAPE but feel there are too many biases associated with MAPE and other percentage errors since my numbers are very small quantities. So am looking for another option to explore testing the accuracy of my model – sarahbarnes Dec 05 '18 at 15:09
  • @sarahbarnes: then you should calculate MAPE error manually based on your training and testing dataset. Instead of ARIMA model MAPE error. – Zeeshan Dec 06 '18 at 05:14