What is Unsupervised learning?

Identification of hidden patterns and associations from data is unsupervised learning

More formally, As compared to Supervised learning where a response variable y is present, in unsupervised learning, there is no response variable y to predict. Thus, there is no explicit mapping from the feature space X to the target Y. Instead the goal is to automatically learn patterns and associations that are hidden in the data.

In some cases, the target is unable to be measured precisely, or it can be unknown entirely. Clustering is one of the primary methods for performing unsupervised learning. This involves clustering observations into categories based on their feature similarities and then deriving and interpreting class labels based on the shared feature values of each grouping.  

Examples:

Clustering
– Clustering population into political ideology based on demographic information.
Please note that the generated clusters will not provide direct information if a group favors one political party over another. Instead, this information has to be inferred.

– Customer segmentation of population based on their purchase history, wage and demographic information.
Again, Clusters will only tell us that a particular group of people belong together. We would have to infer if they are a suitable customer segmentation or not on our own.