Principle:Evidentlyai Evidently Classification Quality Evaluation

Knowledge Sources	A Systematic Analysis of Performance Measures for Classification Tasks Evidently Classification Evidently
Domains	ML_Evaluation, Classification
Last Updated	2026-02-14 12:00 GMT

Overview

A comprehensive classification model quality evaluation mechanism that computes standard performance metrics and optional visualizations.

Description

Classification Quality Evaluation assesses how well a classification model performs by computing standard metrics: accuracy, precision, recall, F1 score, ROC AUC, log loss, and binary-specific rates (TPR, TNR, FPR, FNR). It optionally generates confusion matrices, precision-recall curves, and PR tables.

When reference data is provided, the evaluation includes comparison between current and reference performance, enabling detection of model quality degradation over time.

Usage

Use this principle when evaluating binary or multiclass classification models. Apply it after generating predictions and before reporting results. Requires a DataDefinition with classification task configuration.

Theoretical Basis

Classification quality is measured via standard metrics derived from the confusion matrix:

$Accuracy = \frac{T P + T N}{T P + T N + F P + F N}$

$Precision = \frac{T P}{T P + F P}, Recall = \frac{T P}{T P + F N}$

$F1 = 2 \cdot \frac{Precision \cdot Recall}{Precision + Recall}$

ROC AUC measures the area under the receiver operating characteristic curve, quantifying the model's ability to discriminate between classes across all thresholds.

Related Pages

Implemented By

Implementation:Evidentlyai_Evidently_ClassificationQuality_Preset

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment