Prediction

A prediction is a model’s estimated output for a given input, including a class, value, probability, sequence, or next element.

Plan for this page

Understand one concept in plain language and connect it to nearby ideas.

You will leave with: The word does not always mean forecasting the future. A model “predicts” when it maps a current image to a label or estimates the next token in a sentence.
Time: 3–8 min
Before you begin: No prerequisite. Keep one real example in mind.
Do this now: Read the explanation, then compare the example with the common misunderstanding.

Prepared next stepLearn how inference produces outputs

PredictionA prediction is a model’s estimated output for a given input, including a class, value, probability, sequence, or next element.

Inference
Classification
Regression

Plain explanation

The word does not always mean forecasting the future. A model “predicts” when it maps a current image to a label or estimates the next token in a sentence.

Example

A fraud model may output a 0.08 estimated probability that a particular transaction belongs to the fraud class.

Why it matters

A prediction is an estimate under a model and data context, not a guaranteed fact. Thresholds, calibration, uncertainty, and consequences determine how it should inform action.

Common misunderstanding

Misconception: a prediction of 80 percent means the event must happen. Correction: it expresses an estimate whose quality needs evaluation across comparable cases.

Related concepts

Follow relationships, not only the alphabet

InferenceInference is running a trained model on an input to produce a prediction, generation, score, or other output.ClassificationClassification is assigning an input to one or more defined categories, often with a score or probability for each.RegressionRegression is a modeling task that predicts a numerical value rather than a discrete category.

Learn how inference produces outputs

Sources and scope

What supports this object

Machine Learning GlossaryGoogle for Developers

Technical vocabulary for model predictions; outputs may be scores, probabilities, classes, values, sequences or actions.

Open source

Loading page content...

…