Unsupervised Learning means that we do not have labelled data. In other words, input data samples do not have labeled output. We have input features x1, x2, x3… etc. for the available data samples, but not output y.
With this kind of data, supervised learning algorithms like regression, decision trees etc. are not an option. One can only do Unsupervised Learning with this kind of data. The most popular algorithm in Unsupervised Learning is K-means clustering, which helps separate data into natural classes or natural categories.
The figure depicts the K-means clustering algorithm.