Overview
FeatureByte's SDK Tutorials¶
Welcome to the Credit Default SDK Tutorials.
What You'll Learn¶
This tutorial series walks you through creating a catalog, registering its data model, formulating your use case, crafting features, computing training data, training model and deploying and managing those features. All using FeatureByte SDK.
Dataset Overview¶
In this tutorial, we'll work with the Credit Default Dataset, which provides transactional and historical data for predicting loan defaults. The dataset is composed of of seven tables:
- NEW_APPLICATION: Contains information about new loan applications submitted by clients.
- CLIENT_PROFILE: Describes the profile for each client.
- BUREAU: Lists all previous credits taken by clients from other financial institutions, as reported to the credit bureau.
- PREVIOUS_APPLICATION: Details prior loan applications made by the client.
- INSTALLMENTS_PAYMENTS: Logs monthly installments for loans at the time of payment.
- LOAN_STATUS: Captures the current status of each loan over time.
- CREDIT_CARD_MONTHLY_BALANCE: Provides monthly balance summaries for credit cards previously held by the client with the institution.
Note
If you are interested in a use case that exploits item table, checkout out the Grocery SDK Tutorials.
End-to-End Workflow¶
Define the Data Model of the catalog¶
4. Update descriptions to tables (optional)
5. Set Default Cleaning Operations
Formulate your use case¶
Create features¶
9. Create Window Aggregates from Event Table
11. Create Calendar Window Aggregates from Time Series
Compute training data and train model for your use case¶
13. Compute Historical Feature Values
14. Train LGBM
Deploy your features¶
15. Deploy and serve a feature list
Download the tutorials here¶
Download all the Credit Default Tutorial notebooks here