I am using the forecast package in R to implement ARIMA Models. I am using one variable, the data is from 2016 to 2019 and it is monthly. When I look at the residual check, the 2016 residuals do not look correct. There is a flat like for 2016 and this is having an impact on tuning the model. I am looking for advice on how to fix this issue.


auto.arima output

Residual Plot

fit.arima <- auto.arima(RevenueDataTSTest)


  • $\begingroup$ How are you tuning the model and what exactly is the "impact" on it? $\endgroup$ – whuber Nov 24 '20 at 17:02
  • $\begingroup$ I have not done any tuning at this point. 2016 residuals should not be 0. If I removed 2016, and used 2017 to 2019, 2017 residuals are 0. I'm trying to understand why this is happening. $\endgroup$ – Danielle DeLosa Nov 24 '20 at 18:34
  • 1
    $\begingroup$ The first 12 months are used to initialize the seasonal terms. $\endgroup$ – whuber Nov 24 '20 at 18:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.