Creating a Model Flow

The first step to use machine learning for matching is to create a model.

This task explains how to create a model flow that will be trained and used for matching records within the tenant and matching records from a file with the records in the tenant.
  1. Log in to Console and click Match IQ from the left navigation, or the application list, as highlighted in the following image:

  2. Select the entity type for which you want to create the model and click ADD NEW. For example, if you select the HCP entity type, the model will be used to match HCPs within the tenant and HCPs from an external file with HCPs in the tenant.

  3. Enter a name and description for the model.

  4. Click the down arrow next to the Select the attributes to train the model on field.
    The recommended attributes appear at the top. The recommendations are based on the attributes configured in the match rules defined for the tenant. You can select from the recommended attributes or the any of the other attributes. You need to select attributes that will participate in the matching process to ensure that you get the desired results.

  5. Click DONE.
    The model appears under the Prepare swim lane. Click the entity type (for example, HCP) for which you have created a model flow on the left navigation bar, to view the model flows that are created for the HCP entity type. You will see only one model flow if it is the first flow to be created. At this stage, the records in your tenant are grouped using a process called Blocking. Within each block, records within the tenant are grouped based on the similarities of the attributes you selected.
    Note: Once the activities mentioned above are performed, an email notification is sent to the user who created the model flow. The email notification contains a link using which the user can start training the model.