SSAS data mining with time series

Odpovědět SSAS data mining with time series

  • Wednesday, December 12, 2012 10:43 PM
     
      Has Code

    Here is the context: I have a Time Series table that I have transformed in multiple cases. Each row represents a block of time t for a hockey player. The measures I want to predict using a Neural Network in SSAS are Goals, Assists and GamesPlayed. I also include columns with t-1t-2 and t-3 values additionally to the t value for each of the measures I want to predict.

    Here is the problem: How can I control what do I feed into the inputs and what comes out as the outputs? For instance: I want to feed the data mining model blocks 1 through 12 to predict block 13 and 14. By using the "Save query results" in Mining model prediction tab in SSAS, how could I do this?

    SELECT
      t.[BlockId],
      t.[Player_id],
      [V Time Series Cases].[Games Played T],
      [V Time Series Cases].[Assists T],
      [V Time Series Cases].[Goals T]
    From
      [V Time Series Cases]
    PREDICTION JOIN
      OPENQUERY([Intelli Stats DW],
        'SELECT
          [BlockId],
          [Player_id],
          [Team],
          [Position],
          [OverallDraftRank],
          [Salary],
          [GamesPlayedT],
          [GamesPlayedT-1],
          [GamesPlayedT-2],
          [GamesPlayedT-3],
          [GoalsT],
          [GoalsT-1],
          [GoalsT-2],
          [GoalsT-3],
          [AssistsT],
          [AssistsT-1],
          [AssistsT-2],
          [AssistsT-3]
        FROM
          [dbo].[V_TimeSeriesCases]
        ') AS t
    ON
      [V Time Series Cases].[Team] = t.[Team] AND
      [V Time Series Cases].[Position] = t.[Position] AND
      [V Time Series Cases].[Overall Draft Rank] = t.[OverallDraftRank] AND
      [V Time Series Cases].[Salary] = t.[Salary] AND
      [V Time Series Cases].[Games Played T] = t.[GamesPlayedT] AND
      [V Time Series Cases].[Games Played T-1] = t.[GamesPlayedT-1] AND
      [V Time Series Cases].[Games Played T-2] = t.[GamesPlayedT-2] AND
      [V Time Series Cases].[Games Played T-3] = t.[GamesPlayedT-3] AND
      [V Time Series Cases].[Goals T] = t.[GoalsT] AND
      [V Time Series Cases].[Goals T-1] = t.[GoalsT-1] AND
      [V Time Series Cases].[Goals T-2] = t.[GoalsT-2] AND
      [V Time Series Cases].[Goals T-3] = t.[GoalsT-3] AND
      [V Time Series Cases].[Assists T] = t.[AssistsT] AND
      [V Time Series Cases].[Assists T-1] = t.[AssistsT-1] AND
      [V Time Series Cases].[Assists T-2] = t.[AssistsT-2] AND
      [V Time Series Cases].[Assists T-3] = t.[AssistsT-3]


All Replies

  • Friday, December 28, 2012 8:53 AM
    Moderator
     
     Answered

    Hi HumanGenome,

    You can save the results of a prediction query to a table in any data source that is defined in a Microsoft SQL Server Analysis Services project. You can either create a new table or save the query results to an existing table. If you save the results to an existing table, you can choose to overwrite the data that is currently stored in the table; otherwise, the query results are appended to the existing data in the table. Details about it, please see:

    Mining Model Prediction Tab: How-to Topics: http://technet.microsoft.com/en-us/library/ms175678(v=sql.100).aspx

    Thanks,
    Eileen


    Eileen Zhao
    TechNet Community Support

  • Friday, December 28, 2012 8:18 PM
     
     

    See below :

    http://technet.microsoft.com/en-us/sqlserver/cc510301.aspx


    Ahsan Kabir Please remember to click Mark as Answer and Vote as Helpful on posts that help you. This can be beneficial to other community members reading the thread. http://www.aktechforum.blogspot.com/