AI Dictionary › Fondamenti AI

Classification

Classificazione

Classification is the machine learning task of assigning an observation to a predefined category among a finite set of possibilities, rather than predicting a number as regression does. An email is spam or not spam, an image contains a cat or a dog, a customer will renew or cancel: in all these cases the model chooses among discrete labels.

Definition

How it works

A classifier learns from already-labeled examples (supervised learning) to recognize the patterns that distinguish one class from another. During training it compares its predictions with the true labels and adjusts its parameters to reduce errors. When there are two possible classes it is called binary classification, when there are more than two it is multiclass classification, and when an observation can belong to several categories at once it is multi-label classification.

Applications

It is one of the most widespread tasks in applied AI: image recognition, AI-assisted medical diagnosis, automated content moderation, customer support ticket routing, fraud detection. Almost every family of machine learning algorithms, from decision trees to neural networks, has a variant designed for classification, and it is often the first problem used to benchmark a new model or dataset.

History & etymology

The term comes from the Latin classis, crossed with the scientific practice of sorting objects into categories, a tradition tracing back to eighteenth-century biological taxonomy. In statistics and later in computing, classification took hold in the mid-twentieth century to describe exactly this: assigning items to predefined groups based on their observable characteristics, an idea far older than machine learning but one that machine learning made automatable at scale.

How it's used in Grace

The level you reach in Grace, from Scout to Master, is effectively a classification of your progress into ordered categories: the same principle a classification algorithm applies to data.

Related terms

More in Fondamenti AI

Put it into practice

From our network

Kaimaki Web: Websites That Win Customers

Custom websites, web apps and digital marketing for growing businesses.

Visit kaimakiweb.com →

From the Agora Intelligence blog

More on agora-intelligence.com →

📱 Download the Android app (beta) iOS coming soon

Say what you mean. Get what you need.

Grace Certified, the AI coach that trains and certifies your prompt engineering, by Agora Intelligence.