Matching Models are based on algorithms powered by machine learning.  There are two ways to create a new model.  You can press the ADD MODEL button to develop one from scratch or you can select an existing rule and then press the CLONE button.  In either case, the Model configuration window will be opened which includes these elements:

  • Name - enter an unique name for your rule.
  • Fields - select all the fields that should be included as part of this model.
    Notice that a pencil icon is shown within each field. Pressing this icon will allow you to control how the matching is done for this field. To learn more about this, please follow this link.
  • Continuous Training - If enabled, the new answers will be added to the existing training data (if available). If disabled, the model will be trained from scratch.

Tip:  A cloned model will replicate all the selected fields and other values from the source model but allow you to change them.

When you have completed the configuration, you will have these options:

  • Pressing the TRAIN button will begin the training process.  During this process the model will analyze the data for the fields you specified and select combinations that may or may not be duplicates.  Once a model has completed this process it becomes a matching rule that is will be available for use by datasets.  You can learn about Training here

Note: In some cases, users may wish to train a model on a subset of the data in an object.  To do that, define a dataset with the necessary filters (see Filter in Configuring Datasets: DataGroomr).  If multiple datasets have been created for the current object, the user will be prompted to optionally select a specific dataset.

Good to know : Once training begins the fields for this model will no longer be editable.
  • Pressing the SAVE will only store your changes.  You can return and edit this model.
  • Pressing CANCEL button will exit this window without saving any change.