Unsupervised Learning is used where no classified data is available to the algorithm. The algorithm here has the task of recognizing patterns in the data resulting from clustering of common characteristics.
For example, an unsupervised learning algorithm could be used to search for cancers with similar characteristics to a known patient, based on genetic data.
Predicting the course of diseases
If a corresponding cluster is found, other disease processes in the cluster can be used to predict a course of disease for the respective patient.