« TANAGRA - Part II: Partitioning of dataset into training and testing sets
TANAGRA - Part IV Descriptor selection »

TANAGRA - Part III: Descriptor scaling

Scale the training set

  1. Create a new diagram and configure it to load a training set from a file. This will put a Dataset operator on the diagram.
  2. Put Define status operator (Feature selection) to diagram under the Dataset operator and configure it to set the correct attributes as Input and Target.
  3. Put Standardize operator (Feature construction) to diagram under the Dataset operator and configure it to use the formula (x-x_min)/(x_max-x_min).
  4. Execute.

It is easy to use TANAGRA to scale descriptors in the training set. However, it seems that there is no option to save the parameters used to scale the descriptors in the training set and then apply them on a testing set. This would make it difficult to assess the performance of a model on an independent validation set.

Share This

This entry was posted by Yap Chun Wei on Thursday, April 24th, 2008 at 10:39 pm and is filed under Data mining, Review. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.

3 Responses to “TANAGRA - Part III: Descriptor scaling”

  1. Profnick Says:
    April 25th, 2008 at 6:09 am

    “However, it seems that there is no option to save the parameters used to scale the descriptors in the training set and then apply them on a testing set.”

    This is because in Tanagra the best way of dealing with training/test sets is to keep them all in the dataset but just select and reselect as necessary. So you would carry out your test/train split as you describe but simply deselect the test set when building the model, then reselect when doing the validation. Indeed if you use the “test” option for validation (as opposed to cross validation or LOO) then the default option is to use the unselected data points anyway. Since there is no way of introducing a second data file in TANAGRA, the use of the selection/reselection options is the only way of introducing new data.

  2. Yap Chun Wei Says:
    April 25th, 2008 at 9:50 pm

    Thank you for your explanation.

  3. Sue Mann Says:
    March 18th, 2010 at 2:08 pm

    Hi
    I am doing a PhD on medical scoring systems, using Tanagra. I need to train a set, after randomising, and be able to keep the training rules (which has been developed on the 1st set of randomised data) and to be able to apply them to the testing set. Can you pleas advise me if this is feasible within Tanagra.

    I hope you can help me. I am using Tanagra 1.433
    Thankyou.

Leave a Reply

Close
E-mail It