ARIMA and enough data to work fine


  • I try to do monthly forecast  for sales. But in the data have some products with 2 years of data and another with 1 year almost of data. The problem is with the products that have 1 year datais ARIMA forecast have high results.  What can i do for resolve this issue?  i already try with periodicity hint in 12 and autodetect_periodicity in 0.8 but always have problem with the product that dont have 1 year of data.

    The SQL SERVER is Standard edition.

    Juan Alvarado - MVP SQL SERVER

    Wednesday, November 27, 2013 2:49 AM

All replies

  • Hi Juan,

    Thank you for your question.

    I am trying to involve someone more familiar with this topic for a further look at this issue. Sometime delay might be expected from the job transferring. Your patience is greatly appreciated.

    Thank you for your understanding and support.

    Mike Yin

    If you have any feedback on our support, please click here

    Mike Yin
    TechNet Community Support

    Thursday, November 28, 2013 9:44 AM
  • It is recommended to have at least 3*periodicity (so 3 years) of data to create good ARIMA model.

    It is impossible to create good model on very little data. In your case you might as well use average value or value from the last year as your prediction.

    Tatyana Yakushev []

    Download Predixion Insight 3.0 - World class predictive platform for big data

    Tuesday, December 03, 2013 9:08 AM
  • Hello Juan,

    I am hoping the previous post from Tatyana answers your question. If not, please let me know so I can further research with our experts.


    Meer Al - MSFT

    Monday, December 09, 2013 4:17 AM