models

---

iso_doc_type: "Specification" iso_viewpoint: "ComponentView" type: "module" title: "regression_model_template::core::models — Model Abstraction & Baselines" source_path: "src/regression_model_template/core/models.py" description: "Abstract Model interface with Pydantic validation, Scikit-Learn pipeline wrapper, and SHAP explainability hooks." tags: ["iso42010", "okf", "component_view", "models", "scikit_learn", "shap"] timestamp: "2026-10-01T17:50:00Z" generated: "agent:wiki-iso-documentation-standard" verified: "true" last_verified_commit: "073b5fd"

regression_model_template::core::models — Model Abstraction & Baselines

Source: src/regression_model_template/core/models.py (Lines: L1-L223)
Layer: Domain
Role: Unified model contract ensuring consistent fit, predict, parameter inspection, and SHAP explainability across any regression estimator.

---

1. Class Diagram

classDiagram
    class Model {
        <<abstract>>
        +str KIND
        +get_params(deep: bool = True) Params
        +set_params(**params) Self
        +fit(inputs: Inputs, targets: Targets) Self*
        +predict(inputs: Any) Outputs*
        +explain_model() FeatureImportances*
        +explain_samples(inputs: Inputs) SHAPValues*
        +get_internal_model() Any*
    }

    class BaselineSklearnModel {
        +str KIND = "BaselineSklearnModel"
        +int max_depth
        +int n_estimators
        +int random_state
        -Pipeline _pipeline
        +fit(inputs: Inputs, targets: Targets) BaselineSklearnModel
        +predict(inputs: Any) Outputs
        +explain_model() FeatureImportances
        +explain_samples(inputs: Inputs) SHAPValues
        +get_internal_model() Pipeline
    }
    BaselineSklearnModel --|> Model : implements

Loading

---

2. Method Contracts

BaselineSklearnModel.fit(self, inputs: schemas.Inputs, targets: schemas.Targets) -> 'BaselineSklearnModel'

• Source Citation: src/regression_model_template/core/models.py:L161-L183
• Visibility: Public (+)
• Behavior: Assembles Scikit-Learn preprocessing transformers and Random Forest regressor, fitting estimator on validated inputs and targets.

BaselineSklearnModel.explain_samples(self, inputs: schemas.Inputs) -> schemas.SHAPValues

• Source Citation: src/regression_model_template/core/models.py:L204-L214
• Visibility: Public (+)
• Behavior: Invokes SHAP TreeExplainer on the underlying fitted tree models to calculate local Shapley attribution values.

---

Related: Metrics · Schemas · Training Job