📏  Classification Probability

Classification Probability#

The probability a model assigns to a class for an instance.

Important

✨ AI-generated content. This page was written with the assistance of an AI language model and is provided as a learning aid. Despite careful review, it may still contain mistakes, omissions, or out-of-date information. Whether you are new to the topic, a team lead, or a senior practitioner, treat it as a starting point rather than an authoritative reference: read it critically and independently verify anything you act on (code, commands, figures, and factual claims) against official documentation and primary sources before relying on it.

What it is#

The classification probability is the score a classifier assigns that an instance belongs to a class — an estimate of \(P(\text{class} \mid \text{features})\) between 0 and 1. It is the model’s confidence before any hard decision is made.

From probability to label#

A threshold converts it to a class (in scikit-learn, predict_proba gives the probability, predict applies the cutoff). Two instances scored 0.51 and 0.99 both become “positive,” but they are not equally certain — which is why the probability carries more information than the label.

Why calibration matters#

The probability is only trustworthy if it is calibrated — if events predicted at 0.7 actually happen about 70% of the time. Over- or under-confident scores mislead any downstream risk-based decision, so probabilities are validated with calibration curves, not just accuracy.


Theme: Classification & Averaging Metrics  ·  All terminology



See also

Source article Adapted (context, re-expressed) in our own words from: Classification Probability (insightful-data-lab.com).

Tags: purpose: reference topic: terminology level: intermediate