Introduction to Unsupervised Learning


Machine learning algorithms used to discover patterns in big data that lead to actionable insights and analysis. These different algorithms can be classified into three groups and they are:

Supervised learning

Unsupervised learning

Semi-supervised learning

What is Unsupervised Learning?

Unsupervised learning is the one where we do not know the class label such that as in supervised learning, we have training data or labeled dataset but here we do not have any training sets and the outputs will be depending on the coded algorithm.

Supervised learning problems have divided into two classes like classification and regression, in similar unsupervised learning problems also divided into two. They are:
  • Clustering
  • Association
Clustering: It is an assignment of the set of observations into subsets that are called as clusters so that observations in the same cluster are similar in some manner. It is a common technique for statistical data analysis used in many fields.

This clustering may include various types of clustering like Hierarchical clustering, k-means clustering, Gaussian mixture models, Self-organizing maps, Hidden Markov models.

Association: It is a method for discovering interesting relations between variables in large databases.

Apriori algorithm is the best example to show how the association works.

Applications of Unsupervised Learning:

a) Neural Networks.

b) Bioinformatics for sequence analysis

c) Genetic clustering

d) Sequence and pattern finding (Data Mining)

e) Image segmentation

f) Object recognition (Computer Vision)