Unsupervised Learning

In short

A type of Training where you don’t know the solution — you just have the task at hand and the goal is to form some kind of structure or grouping from it.

Like sorting a pile of photos without anyone telling you what the categories should be. You’d naturally start grouping them — family, vacations, work events — based on patterns you notice.

Where Supervised Learning requires labeled Data (inputs paired with known answers), unsupervised learning works with raw, unlabeled data — the Model has to figure out the structure on its own. An important use case is in marketplaces. You want to understand your customers and how they behave, so you can make better deals or proposals to stimulate buying. You take the data you collected — what each user bought, when, how much — and feed it into a model that creates groups of clients based on their activity. Nobody told the model what the groups should be; it found them on its own.

Related

• Training - unsupervised learning is a type of training
• Supervised Learning - the other main approach
• Data - needs data, but no labels/solutions required