Implementation:DistrictDataLabs Yellowbrick PrePredict

Knowledge Sources	DistrictDataLabs_Yellowbrick
Domains	Model_Evaluation, Utilities
Last Updated	2026-02-08 05:00 GMT

Overview

Passthrough estimator that wraps pre-computed predictions for use with Yellowbrick visualizers without re-running inference.

Description

The PrePredict class implements the scikit-learn estimator interface but returns pre-computed predictions instead of running a model. This is useful when model inference is expensive (e.g., deep learning models) and results have already been cached. It supports classification, regression, and clustering metrics via automatic estimator type detection.

Usage

Import PrePredict when you have pre-computed model predictions and want to use Yellowbrick visualizers without re-running inference. Wrap the predictions in a PrePredict object and pass it as the estimator to any Yellowbrick visualizer.

Code Reference

Source Location

• Repository: DistrictDataLabs_Yellowbrick
• File: yellowbrick/contrib/prepredict.py
• Lines: 1-104

Signature

class PrePredict(BaseEstimator):
    def __init__(self, data, estimator_type=None):
        """Passthrough estimator for pre-computed predictions.

        Parameters
        ----------
        data : array-like, callable, or path
            Pre-computed predictions, a callable returning them, or a file path.
        estimator_type : str, optional
            One of "classifier", "regressor", or "clusterer".
        """

    def fit(self, X, y=None): ...
    def predict(self, X): ...
    def score(self, X, y=None): ...

Import

from yellowbrick.contrib.prepredict import PrePredict

I/O Contract

Inputs

Name	Type	Required	Description
data	array-like, callable, or path	Yes	Pre-computed predictions
estimator_type	str	No	"classifier", "regressor", or "clusterer"

Outputs

Name	Type	Description
predict()	array-like	Returns the pre-computed data
score()	float	Metric score (accuracy, R2, or silhouette)

Usage Examples

import numpy as np
from yellowbrick.contrib.prepredict import PrePredict
from yellowbrick.classifier import ClassificationReport

# Pre-computed predictions
y_pred = np.array([0, 1, 1, 0, 1, 0, 0, 1])
y_true = np.array([0, 1, 0, 0, 1, 1, 0, 1])

model = PrePredict(y_pred, estimator_type="classifier")
viz = ClassificationReport(model)
viz.score(None, y_true)
viz.show()

Related Pages

• Environment:DistrictDataLabs_Yellowbrick_Python_Scikit_Learn_Environment