Time Series algorithm question

Discussion in 'microsoft.public.sqlserver.datamining' started by Tim, Aug 16, 2006.

  1. Tim

    Tim Guest

    Trying to create very easy TS model

    Month long Time Key
    Revenue double predict
    Person text Key

    I have a problem to interpret the results


    1. In Mining Model Viewer Number of forecast months is less that Prediction
    steps.
    The same result I get through DMX- just 3 months of data, others are empty .
    2. Historical data on vertical axis (Abs or Procentage) are too different to
    SELECT Month, SUM(Revenue) FROM........GROUP BY Month
    How to get real historical abd forecast data?

    1. What format for data (long ) field - Iam using 200601, 200602....etc ,
    but TS generate dates like 200614, 200615,,,
    Should I use DATE data type instead?

    Thanks,
    Tim.
     
    Tim, Aug 16, 2006
    #1
    1. Advertisements

  2. The TS algorithm stops prediction if it thinks the series is unstable. In
    later versions we will have parameters to control this. There is no time
    intelligence built into the key time - it is simply a time stamp indicating
    order of data points.

    --

    -Jamie MacLennan
    SQL Server Data Mining
    This posting is provided "AS IS" with no warranties, and confers no rights.
    PLEASE POST NEW QUESTIONS AT
    http://forums.microsoft.com/msdn/showforum.aspx?forumid=81&siteid=1
    PLEASE POST NEW QUESTIONS AT
    http://forums.microsoft.com/msdn/showforum.aspx?forumid=81&siteid=1
    PLEASE POST NEW QUESTIONS AT
    http://forums.microsoft.com/msdn/showforum.aspx?forumid=81&siteid=1
    PLEASE POST NEW QUESTIONS AT
    http://forums.microsoft.com/msdn/showforum.aspx?forumid=81&siteid=1
     
    Jamie MacLennan \(MS\), Aug 17, 2006
    #2
    1. Advertisements

  3. Tim

    Tim Guest

    Jamie,
    Thank you for prompt respond.
    These is one more problem which bother me most becauseI deal with $$$.
    So, trying to predict Revenue gererated by each Person.
    My Input like this:
    200501 Person1 $1000
    200501 Person1 $2000
    200502 Person1 $1000
    200501 Person2 $5000
    200502 Person2 $2000
    200502 Person2 $3000

    Obviosly for Person1 and 200501 I expect to see $3000, correct?
    Instead on MS Time Series Viewer I see REVENUE(actual) - 200501 VALUE =XXX,
    Where XXX is absolutly different number.

    Person1 who is tough guy tryed to shoot me.
    What I am doing wrong. Could you please give me an idea who to extract real
    historical and predict information?

    Thnak you,
    Tim.
     
    Tim, Aug 17, 2006
    #3
  4. Tim

    Tim Guest

    Jamie,
    Thank you for prompt respond.
    These is one more problem which bother me most becauseI deal with $$$.
    So, trying to predict Revenue gererated by each Person.
    My Input like this:
    200501 Person1 $1000
    200501 Person1 $2000
    200502 Person1 $1000
    200501 Person2 $5000
    200502 Person2 $2000
    200502 Person2 $3000

    Obviosly for Person1 and 200501 I expect to see $3000, correct?
    Instead on MS Time Series Viewer I see REVENUE(actual) - 200501 VALUE =XXX,
    Where XXX is absolutly different number.

    Person1 who is tough guy tryed to shoot me.
    What I am doing wrong. Could you please give me an idea who to extract real
    historical and predict information?

    Thnak you,
    Tim.
     
    Tim, Aug 17, 2006
    #4
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.