Quest Modeler

The Quest Modeler is the tool for building a trained model. You can use the drag-and-drop visual panel to assemble, connect, and configure a flow of linked actions. These are the machine learning phases.

Creating a model requires these steps:

  1. Import the data
  2. Define preprocessing steps
  3. Apply an algorithm
  4. Train the model
  5. Score and evaluate

Activities are available for selection in a catalog format in the Quest Modeler. The catalog is displayed in a popup window and you can add the desired activity by moving the activity to the canvas. You can filter the activities by category.

In the Quest Modeler, each activity is represented as a box. When a box is selected, a respective configuration panel is displayed in the right hand-side. Specify a name and description for the activity and configure the parameter options for the activity.

Draw an arrow from the output port of one activity to the input port of another activity to define the sequence.

These are the two phases in the model construction:

  • Training
  • Production