Administration > Relationship discovery > Relationship inference model
  

Relationship inference model

The relationship inference model is a predefined model consisting of user-defined rules and machine learning components. You can use this model to discover the relationships between columns within a catalog source.
The relationship discovery capability infers table-level relationships on the basis of column similarity relationships.
Metadata Command Center provides a predefined relationship inference model called Column Similarity Model v1.0. This model discovers similar columns within a catalog source based on the similarity among column values. You can import this model and associate it with any catalog source to discover the relationships defined in the model. For information about importing predefined content, see the Introduction and Getting Started help.
To associate a relationship inference model with a catalog source, you must enable the Relationship Discovery option and specify parameters while creating a catalog source. For more information about creating a catalog source, see the Catalog Source Configuration help.

Import a relationship inference model

Metadata Command Center allows you to import a relationship inference model to infer column-level and table-level relationships within a catalog source.
To import a relation inference model, perform the following steps:
    1In Metadata Command Center, go to the Explore page in the left navigation panel.
    2Click the arrow () icon on the top of the page and select Relationship Inference Model.
    3On the Relationship Inference Model window, click Import.
    The Import Relationship Inference Model dialog box appears.
    Image of the Import Relationship Inference Model dialog box
    4Click Choose File to navigate to the predefined model file on your local machine. You can also drag and drop the file.
    5Click OK.
    The imported model appears in the list of models on the Relationship Inference Model window.
After you import the relationship inference model, you can use the model to identify pairs of similar columns and table level relationships in a catalog source. You can associate the relationship inference model with a catalog source while creating a catalog source. For more information about creating a catalog source, see the Catalog Source Configuration help.