Hitachi Vantara Pentaho Community Forums
Results 1 to 2 of 2

Thread: WekaForecaster - buildForecaster Vs primeForecaster

  1. #1

    Default WekaForecaster - buildForecaster Vs primeForecaster

    If we have built our forecaster(model) using buildForecaster() method, why we have to call primeForecaster() (on the same training data) before calling the forecast() method?

  2. #2
    Join Date
    Aug 2006
    Posts
    1,741

    Default

    WekaForecaster uses standard propositional machine learning regression methods under the hood to make forecasts. In order to remove the time dependency between data points it transforms it into independent instances by encoding the time dependency into "lagged" versions of the target variable to predict. This is done for the training data, and then the underlying regression model is built. In order to create a transformed instance as input to generate a forecast, it is necessary to see enough historical instances in order to populate the values of the lagged variables. If you use a maximum lag length of 12, then it is necessary to input the most recent 12 historical data points in order to create one transformed instance that has the values "flattened out" into the lagged variables - I've called this process "priming" (for lack of a better term).

    Cheers,
    Mark.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  
Privacy Policy | Legal Notices | Safe Harbor Privacy Policy

Copyright © 2005 - 2019 Hitachi Vantara Corporation. All Rights Reserved.