# Retrain Time Series Models

I'm new to TS modeling, but have some experience in classic classification modeling. In classification I can train one model and use it for some time while some indices are stable (e.g. PSI).

Assuming I want to forecast daily demand on bread. One day (31 DEC 2013) I've trained ARIMA model and made a forecast for tomorrow (1 JAN 2014). Is there any way to use this model tomorrow (1 JAN 2014) without retrain it on new data (demand before 1 JAN 2014 and demand on 1 JAN 2014)?

The reason to ask: in case of huge amount of goods it requires a lot of time to retrain all models everyday. Retrain it once a month (like in classification usage pattern) will be better.

• This question seems more suited for Stack Overflow, or is there any methodological issue in your question? – javlacalle Nov 17 '14 at 10:31
• @javlacalle this question is about methodology of TS-model usage. If there are more (or only) technical solutions I will be glad to see this question on SoF. – Aleksandro M Granda Nov 17 '14 at 10:46

The function Arima in the forecast allows you to apply a model from a previous call to new data without re-estimating the model. If this is what you are looking for, you can just pass the output from a previous fit through the argument model.