6. Formulate Use Case

Why Create a Use Case?

Creating a Use Case in FeatureByte, while optional, unlocks a range of powerful benefits. It serves as a clear blueprint for your prediction data, outlining its purpose and intended application. Moreover, Feature Ideation leverages this information to suggest features specifically tailored to your use case, streamlining feature engineering.

Let's Create a Use Case for Our Credit Default Dataset.

Loan Default by client: Predict clients' payment difficulties the next 6 months for new loan based on its early stage application data and prior credit history.

---

Step 1: Create a Context

A Context defines the scope and conditions in which features will be served.

Steps to add a Context

1. Navigate to the Context catalog in the 'Formulate' section of the menu.
2. Click

3. Define the Context for this use case:

   • Context Name: New Loan Application (Early Stage)
   • Description: Early-stage loan application under review, without external data.
   • Primary Entity: New Application

After adding the context, it appears in the Catalog:

---

Step 2: Establish a Target

In FeatureByte, you can set a target in two ways:

• Logical Approach: Similar to feature creation, this approach allows the computation of targets through FeatureByte.
• Descriptive Approach: Simply state what you want to predict.

For this use case, we'll use the descriptive approach, as the target cannot be derived directly from the data and must be provided alongside the observation tables.

Check out

For an example of the logical approach, check out the Grocery UI Tutorial.

Steps to add a Target

1. Navigate to the Target catalog in the 'Formulate' section of the menu.
2. Click

3. Define the target using the descriptive approach:

   • Target Name: Loan_Default
   • Primary Entity: New Application
   • Data Type: Integer
   • Description: Whether the loan will default the next 6 months.
   • Target Type: Classification
   • Include Horizon: 182 days.

After adding the target, it appears in the Catalog. The sign indicates that the target cannot be computed automatically and must be provided alongside the observation tables.

---

Step 3: Create a Use Case

Steps to add a Use Case

1. Navigate to the Use Case catalog in the 'Formulate' section of the menu.
2. Click

3. Define the use case as follows:

   • Use Case Name: Loan Default by client
   • Primary Entity: New Application
   • Context: New Loan Application (Early Stage)
   • Target: Loan_Default
   • Description: Predict clients' payment difficulties the next 6 months for new loan based on its early stage application data and prior credit history.

Note

Detailed use case descriptions enhance FeatureByte's capability to provide relevant feature suggestions.

Step 4: Review a Use Case

Once created, the use case appears in the Catalog:

Click on the 'Loan Default by client' use case to review it.