Skip to content

Use Case Gallery

Welcome to the VerifIA Use Case Gallery—four end‑to‑end examples demonstrating how VerifIA can be applied to both regression and classification problems. Each use case includes a Jupyter notebook that guides you through:

  • Wrapping a trained model
  • Loading a predefined domain YAML or generating one via AI
  • Running rule‑consistency verification
  • Interpreting and exporting results

By exploring these, you’ll see how VerifIA adapts across data types, frameworks, and verification goals.

Regression Use Cases

1. Compressive Strength Regression (TensorFlow)

Verify a TensorFlow regression model trained on concrete compressive strength data. This guide demonstrates:

  • Hyperparameter tuning with scikit-optimize
  • Manual & AI‑powered domain generation
  • Rule‑based verification with VerifIA

Open in Colab

Artifacts:

2. House Pricing Prediction (CatBoost)

Verify a CatBoost regression model forecasting house prices. You will learn:

  • Bayesian tuning with BayesSearchCV
  • Wrapping CatBoost via CBModel
  • Loading or generating a domain config
  • Verifying rule‑consistency with VerifIA

Open in Colab

Artifacts:

Classification Use Cases

3. Hotel Cancellation Prediction (Scikit‑Learn)

Verify a Scikit‑Learn pipeline (SVC & Random Forest) for predicting hotel cancellations. This walkthrough covers:

  • Building preprocessing pipelines (StandardScaler, OneHotEncoder)
  • Tuning with RandomizedSearchCV & BayesSearchCV
  • Wrapping via SKLearnModel
  • Domain config & rule‑consistency verification

Open in Colab

Artifacts:

4. Loan Eligibility Classification (XGBoost)

Verify an XGBoost classifier for loan repayment prediction. In this example you will:

  • Perform Bayesian hyperparameter tuning (BayesSearchCV)
  • Wrap using XGBModel
  • Generate or load domain configurations
  • Run rule‑consistency checks

Open in Colab

Artifacts:

Why Use Cases?

  • Hands‑on guidance: Detailed notebooks covering each step.
  • Framework diversity: TensorFlow, CatBoost, Scikit‑Learn, XGBoost.
  • Balanced scenarios: Two regression and two classification examples.
  • Best practices: Tips on tuning, domain setup, and result interpretation.

Next Steps:
1. Choose the use case matching your problem type.
2. Clone the corresponding notebook and run it locally.
3. Adapt the patterns to verify your own models with VerifIA.